- Email: [email protected]
Set-up and calibration of a method to measure 10B concentration in biological samples by neutron autoradiography
Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
Set-up and calibration of a method to measure 10B concentration in biological samples by neutron autoradiography M.A. Gadan a,b, S. Bortolussi b,c,⇑, I. Postuma b, F. Ballarini b,c, P. Bruschi b, N. Protti b,c, D. Santoro b,c, S. Stella b,c, L. Cansolino d, A. Clerici d, C. Ferrari d, A. Zonta d, C. Zonta d, S. Altieri b,c a
National Commission for Atomic Energy (CNEA), Buenos Aires, Argentina Department of Nuclear and Theoretical Physics, University of Pavia, Pavia, Italy National Institute for Nuclear Physics (INFN), Section of Pavia, Pavia, Italy d Department of Experimental Surgery, University of Pavia, Pavia, Italy b c
a r t i c l e
i n f o
Article history: Received 19 October 2011 Received in revised form 24 November 2011 Available online 8 December 2011 Keywords: Boron neutron capture therapy Boron concentration measurements Neutron autoradiography CR-39
a b s t r a c t A selective uptake of boron in the tumor is the base of Boron Neutron Capture Therapy, which can destroy the tumor substantially sparing the normal tissue. In order to deliver a lethal dose to the tumor, keeping the dose absorbed by normal tissues below the tolerance level, it is mandatory to know the 10B concentration present in each kind of tissue at the moment of irradiation. This work presents the calibration procedure adopted for a boron concentration measurement method based on neutron autoradiography, where biological samples are deposited on sensitive films and irradiated in the thermal column of the TRIGA reactor (University of Pavia). The latent tracks produced in the film by the charged particles coming from the neutron capture in 10B are made visible by a proper etching, allowing the measurement of the track density. A calibration procedure with standard samples provides curves of track density as a function of boron concentration, to be used in the measurement of biological samples. In this paper, the bulk etch rate parameter and the calibration curves obtained for both liquid samples and biological tissues with known boron concentration are presented. A bulk etch rate value of (1.64 ± 0.02) lm/h and a linear dependence with etching time were found. The plots representing the track density versus the boron concentration in a range between 5 and 50 lg/g (ppm) are linear, with an angular coefficient of (1.614 ± 0.169)103 tracks/(lm2 ppm) for liquids and (1.598 ± 0.097)102 tracks/(lm2 ppm) for tissues. Ó 2011 Elsevier B.V. All rights reserved.
1. Introduction Boron Neutron Capture Therapy (BNCT) is an experimental form of radiotherapy consisting of two phases: administration of a 10B carrier able to concentrate this element in tumor more than in healthy tissues, and irradiation of the target with low energy neutrons [1]. The high LET radiation originating from neutron capture in 10B, irreversibly damages the cell in which the reaction takes place. If a sufficient boron concentration can be loaded in tumor cells and a low boron concentration is absorbed by normal tissues, the irradiation can lead to the selective destruction of essentially all the tumor cells present in the target. In order to assess a proper treatment plan, that is to deliver an effective irradiation with a substantial sparing of the healthy cells, it is necessary to know the boron concentration in each kind of irradiated tissue. In BNCT research, many methods are used to measure boron concentration in biological samples. One of the ⇑ Corresponding author. Address: Silva Bortolussi, Dip. Fisica Nucleare e Teorica, Università di Pavia, via Bassi 6, 27100 Pavia, Italia. Tel.: +39 0382 987635. E-mail address: [email protected] (S. Bortolussi). 0168-583X/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2011.11.043
most common ones is Prompt Gamma Neutron Activation Analysis (PGNAA) [2,3], a radio-analytic technique which exploits the 0.478 MeV c ray that is emitted from the excited 7Li in 93.9% of the capture reactions. With a facility dedicated to BNCT, it is usually possible to measure few ppms of 10B in relatively large samples (0.5 cm3) with irradiation times of some minutes. Of course, to measure low boron concentrations in small samples it is necessary to increase the measurement time. This technique is limited by the relatively high dimensions and the heterogeneity of the analyzed samples. In this case, the concentration that is obtained from the measurement procedure is an average value that cannot represent the uptake of the various cell fractions of the sample, such as viable tumor cells, healthy cells, necrosis, fibrosis and so on. The same problem is shared by other measurement methods such as ICP-AES (Inductively-Coupled Plasma-Atomic Emission Spectrometry) [4–7], which allows measuring concentrations in the range 0.2–70 ng/g. A more precise technique from the point of view of the spatial resolution is SIMS (Secondary Ion Mass Spectrometry) [5], which is a particular form of mass spectrometry. The sputtering of the sample surface with a beam of primary ions produces the emission
52
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
of ions from the sample. These ions are separated according to their mass, and their signals are transformed in an electronic map that constitutes a bi-dimensional image of the sample surface. The quantity of sample necessary for this technique is very small if the ion beam is focused in spots of 0.05–0.5 lm; boron concentrations that can be measured are in the interval 0.1–10 ppm. The Boron concentration measurement method used in Pavia for biological tissues is based on the spectroscopy of the charged particles coming from the reaction 10B(n,a)7Li, when irradiating thin tissue sections in the thermal column of TRIGA reactor [6,7]. The sections are laid on mylar supports and irradiated in a vacuum chamber facing a Si detector. This method is not suitable for liquid samples, like blood and urine, because they cannot stay under vacuum. To solve this problem, a method based on neutron autoradiography is proposed. The autoradiography technique is based on the interaction of high-LET particles crossing a material and producing damages in its molecular structure along the particle tracks. Since in conductors and semiconductor materials a recombination process occurs only dielectric materials show regions with an irreversible damage. These regions are called latent tracks. By exposing the material to a proper etchant solution, chemical reactions occur everywhere in the material, but more intensively in the latent track regions. This effect of the etching process produces track pits on the surface of the material, which are visible with an optical microscope. Because of their properties, dielectric materials such as plastics and glasses are used in the autoradiography technique, such as sensitive supports called SSNTDs (Solid-State Nuclear Track Detectors) [10]. In the case of boron concentration measurements in biological samples, a small amount of sample is deposited on a SSNTD, for the irradiation in the neutron field. The SSNTD is then etched with a proper solution to visualize the damage produced by high-LET particles during the sample–neutron interaction. Finally, the images taken with an optical microscope can be analyzed with a suitable software to perform different kinds of analysis. Depending on the neutron fluence, the autoradiography technique can be employed to make quantitative or qualitative analysis. On one hand, at a high neutron fluence, a higher density of high-LET particles coming from 10B(n,a)7Li reaction is produced, and with a proper etching it is possible to obtain images for macroscopic boron bio-distribution qualitative analysis. A method to study the boron bio-distribution in biological samples by autoradiography imaging was already assessed in Pavia [8,9]. The irradiation of the samples and the etching parameters were optimized to obtain images of 10B distribution, which were compared to the histological preparation of a contiguous tissue section. In this way it is possible to verify if the areas with higher track density correspond to the tumor, and thus if the tumor cells are loaded with higher boron concentrations. On the other hand, irradiating with a low neutron fluence, the density of high-LET particles produced is lower, thus causing less tracks in the detector. A proper etching of the SSNTD allows visualizing the tracks separately, and counting them. Images are collected with an optical microscope and are analyzed with an adequate software to determine the track density, which is used to calculate the boron concentration in the sample through a proper calibration curve. The main aim of the present work was to measure the calibration curve for liquid and solid biological samples. The paper describes the experimental set-ups, the preparation of the calibration samples, and the results of the measurements.
2. Materials and methods The sensitive films used for neutron autoradiography were rectangular polyallyldiglycol carbonate (PADC) CR-39 film
detectors from Intercast Europe manufacturer, with a 75 25 mm2 area and a 1 mm thickness. Before starting with boron concentration measurements, the etching process in CR-39 films was characterized by the determination of the bulk etch rate parameter, which is useful to compare the results of autoradiographies obtained under different etching conditions. Another important parameter is the neutron flux at which the samples were exposed, in order to set the fluence received by the samples with good accuracy. To characterize the calibration setup, the neutron flux was measured in the facilities used for liquid and solid biological sample irradiation. Calibration curves were assessed both for liquid and for solid biological samples in a range between 5 and 50 ppm. 2.1. Bulk etch rate measurement The formation of a track depends on the competitive effect of two parameters: the track etch rate (VT) and the bulk etch rate (VB). The track etch rate is the velocity at which the etching process progresses along the latent track, whereas the bulk etch rate is the etching velocity at undamaged regions of the detector. Considering a particle with normal incidence on the detector, and a constant track etch rate (VT) and bulk etch rate (VB), the diameter of the track (D) at an etching time (t) can be expressed as follows [10]:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi VT VB D ¼ 2 VB t VT þ VB
ð1Þ
Eq. (1) shows the competitive effect between VT and VB in the growth of the track diameter. Since for heavy ions and fission fragments VT VB, the removed layer, and thus VB, can be related to the track opening diameter by the relation [10]:
DF VB t 2
ð2Þ
where DF is fission fragment track diameter, VB is bulk etch rate and t is etching time. For VB measurement, the CR-39 films were irradiated with a 252 Cf source at a short distance and etched for the following times: 30, 45, 60, 70 and 80 min, in a 6.25 N NaOH solution at 70 °C. The etching solution was prepared mixing 99.0% purity NaOH with pure water. After etching, CR-39 films were washed out with cold water in order to remove the etchant solution from the detector surface. The experimental set up for the image acquisition is composed by a stereomicroscope (Leica MZ16A) connected to a lamp (Leica CLS150X) and a joystick (Prior Optiscan II). The microscope has an integrated camera connected to a PC and the software LAS V3.7 was employed to acquire the pictures. The analysis of fission fragment tracks was performed with the ImagePro Plus Analizer 6.3 code. 2.2. The irradiation facilities and the neutron flux characterization Liquid samples were irradiated in the facility shown in Fig. 1, which allowed pouring 5 ml of solution in sealed holes below which CR-39 films were positioned. This device hosted four CR-39 films, and for each film four holes for liquid samples were available. In this way it was possible to obtain films in which up to four different concentrations were simultaneously present. The neutron flux was measured by means of copper foil activation using the Wetscott method [11]. Copper foils were positioned inside four holes, one in each CR-39, filled with water and irradiated in the same position as for the real measurements (see Section 2.3).
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
53
Fig. 1. Liquid samples irradiation facility. It is visible the inferior plexiglass bar with 4 housings for CR-39 slides and the superior one with the sealed holes where the boronated solutions were poured.
Fig. 2. Picture of the etched CR-39, for a liquid sample with 50 ppm of irradiated with a fluence of (5.18 ± 0.21) 1010 n/cm2.
After irradiation, the foils were taken off the facility and their activity was measured by a germanium detector. Solid biological samples were irradiated in a facility made up of two plexiglass slides positioned vertically at the end of the thermal column. This facility allows the irradiation of up to thirty CR-39 films positioned at different heights in the same plane. The neutron flux determination was carried out using the same method as for liquid irradiation facility, positioning copper foils in representative points of the CR-39 irradiation facility.
parameters were set to obtain the highest contrast between the background and the interesting points of the picture. A shadow correction (corresponding to a background subtraction) was applied to remove the inhomogeneous light field, which would alter the quality of the pictures. Once the images were acquired, the code Image-Pro 6.3 was used to classify the charged particle tracks, visualized as dark objects on a uniform gray background (Fig. 2). The analysis process required the binarization of the images, using the gray scale histogram. A built-in code was then employed to calculate some userdefined parameters, such as the minimum diameter of the tracks and their area. These two parameters where used to select the interesting points to be counted. The software employed for the analysis was a home-made script written in the data analysis framework ROOT [12]. The distribution of the track area was used to set a threshold corresponding to the objects of 1 pixel, which are not tracks but artifacts that the binarization program could not delete. Another threshold was set for too large objects, which could be multiple tracks, or artifacts, or tracks not produced by a particles. These events are quite unusual, but they must be excluded from the counts. The minimum diameter distribution was used to define the range of diameters where the objects to be counted lay. To this aim, all the photos taken for each concentration value were used to plot a histogram of the minimum diameters, and this distribution was plotted with a Gaussian curve (Fig. 3). The interval [ld 3rd, ld + 3rd] was chosen for each concentration value as the range where the tracks must be counted. In each picture, the tracks within this interval were then counted and plotted as a histogram and fitted with a Gaussian curve (Fig. 4). The mean value lc of this fit was divided by the area of the field and used as mean track density for the considered concentration; 1r was used as the error associated to this point. Finally, the points obtained for all the concentration values were plotted as a function of boron concentration, and the linear fit was computed.
2.3. Preparation and irradiation of the calibration samples Liquid standards consisting of water solutions of boric acid at boron concentrations of 5, 10, 25 and 50 ppm were poured on CR-39 films, using the device constructed on purpose for the irradiation of liquid samples. Liquid samples were irradiated at the end of the thermal column of the TRIGA Mark II reactor, operating at 10 kW for 30 min, receiving a neutron fluence of (5.18 ± 0.21) 1010 n/cm2. After irradiation, the CR-39 films were chemically etched for 60 min with the same etching parameters used for the fission fragment measurement. Samples obtained mixing cell suspensions of rat healthy liver with BPA-HCl, at different boron concentrations, were used for tissue calibration. Briefly, cell suspensions were obtained by mechanically treating the liver tissue with a potter homogenizer. The tissue inserted into the tube was pressed until its complete disaggregation. Then the obtained cell suspensions were divided into five aliquots, enriched by centrifugation and weighted. A boronophenylalanine–HCl (BPA–HCl) solution at 500 ppm of 10B was added to each cell fraction in order to obtain the reference boron concentrations of 5, 10, 25 and 50 ppm. The concentration values were then recalculated on the basis of the true weights of each prepared sample; the error associated to the concentration values is of the order of 0.5%. Finally, the samples were frozen in small cylindrical rods at 80 °C. Slices of 60 lm were obtained using a Leica cryostat and were deposited on CR-39 films, which were then irradiated at the end of the thermal column with the reactor operating at 5 kW for 30 min, receiving a neutron fluence of (4.91 ± 0.28) 1010 n/cm2. After irradiation, the films were chemically etched at the described etching conditions for 45 min. 2.4. Track density analysis After the etching process, a number of images from the irradiated samples were taken with the stereomicroscope. In order to measure the track density with good accuracy, acquisition
10
B,
2.5. The dry to fresh mass ratio of the samples When using the calibration curve obtained for solid samples, an important parameter must be taken into account: the amount of water loss occurring after the preparation of the tissue sections. Tissues with the same composition as the calibration samples could lose water in different percentages. Therefore, when calculating boron concentration in a sample by comparison with the calibration curve, the results must be renormalized by the ratio between the water loss in the calibration samples and in the tissue
54
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
Fig. 3. Example of histogram representing the minimum diameter of the tracks, obtained from all the pictures taken for a single concentration value.
Fig. 5. Bulk etch rate curve.
achieved in the track diameter measurement, the relative error decreases with the etching time, and it is maximum (13%) for an etching time of 30 min. 3.2. Neutron flux measurements in the facilities The thermal neutron flux measured in the liquid irradiation facility with the reactor operating at 250 kW was (7.20 ± 0.29)108 n/cm2s, thus in the liquid irradiation position there was a flux of (2.88 ± 0.12)107 n/cm2s with the reactor operating at 10 kW. The thermal neutron flux measured in the solid sample irradiation facility with the reactor operating at 250 kW was (1.37 ± 0.08)109 n/cm2s and the flux in the plexiglass holding at a power of 5 kW was (2.74 ± 0.16)107 n/cm2s. 3.3. Calibration for liquid samples Fig. 4. Example of a histogram representing the number of tracks/picture obtained from all the pictures taken for a single concentration value. The mean value of the Gaussian, divided by the area of the picture, was used in the calibration curve; the sigma of this distribution is the error associated to the point.
sample. In order to measure the ratio between the weight of a dry calibration sample and the fresh one, a number of 60 lm sections were cut and weighted by a high sensitivity digital scale (up to 1 lg). Since loss of weight occurs very quickly for thin sections, the scale was connected to a Personal Computer and the weight was acquired every second by means of a home-made acquisition software. When the section was completely dried, that is the plot of the weight versus time was constant, the dry to fresh weight ratio was calculated. The process was repeated several times, and the ratio was obtained as an average of all the performed measurements.
The calibration curve for liquid samples is shown in Fig. 6. The fit of the plot is linear, with an angular coefficient of (1.614 ± 0.169)103 tracks/(lm2 ppm). The v2/ndf of the fit is 0.9/3. Taking into account a VB of 1.63 lm/h and an etching time of 60 min, the removed layer was (1.64 ± 0.02) lm. The errors associated with the points at low concentrations are quite high, depending on the low track density in the pictures. This could be avoided irradiating with higher neutron fluence. Nevertheless, this experiment was conceived to measure the track density in a range between 5 and
3. Results 3.1. Bulk etch rate The thickness of the removed layer as a function of the etching time according to (2) is shown in Fig. 5. The track diameter increases with the etching time for the whole considered time range, and a linear response with a constant bulk etch rate of VB = (1.64 ± 0.02) lm/h was observed. This value is in good agreement with those reported in the literature for fission fragments and the same etching parameters [13]. Thanks to the precision
Fig. 6. Calibration curve for liquid samples; the angular coefficient is indicated as p1.
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
50 ppms, and the fluence was chosen in a way that the tracks remained separated also in the samples with higher boron concentration. When samples with low concentration are measured, it would be always possible to make another calibration for a small range of boron concentrations, irradiating with a higher neutron fluence and thus lowering the errors. 3.4. Calibration for tissue samples The calibration curve for tissue samples is shown in Fig. 7. Again, the fit is linear, with an angular coefficient of (1.598 ± 0.097)102 tracks/(lm2 ppm). The v2/ndf of the fit is 2.6/3. For an etching time of 45 min the removed layer was (1.23 ± 0.01) lm. 3.5. The dry to fresh mass ratio of the samples The dry to fresh weight ratio measured for the frozen cell suspension was (0.15 ± 0.01). Fig. 8 shows the weight loss by a tissue calibration sample. Experimental measurements showed that, after 1500–1800 s, the variation in the weight loss was less than 1%, under our laboratory conditions. The rise in the first part of the curve is due to the sample position in the scale. After this steep rise, the tissue sample weight decreases down to a limit value. The ratio between this limit value and the fresh weight is the parameter to be used in the concentration calculation. 4. Discussion and conclusions Calibration curves obtained both for liquid and for solid samples showed a very good linearity in the range between 5 and 50 ppm, allowing boron concentration measurements in biological samples,
55
both liquid and solid, for BNCT purposes using neutron autoradiography in the described conditions. It is very important to note that a proper calibration is needed for each kind of sample that must be measured. In fact, the stopping power of the charged particles depends on the material that they cross. For this reason, the calibration for solid biological tissues was performed using liver tissue reduced to a cell suspension. In this way, when tissue samples are cut, deposited on CR-39 and irradiated, the obtained track density can be related to the calibration curve, because the stopping power of a particles is the same. Therefore, it is important to use tissue calibration when measuring tissues, water solution calibrations when measuring liquids, and so on. Furthermore, it must be taken into account that the sectioned tissues dry very rapidly, modifying their water content. Our experience shows that the biological variability of this weight change is high and each type of tissue must be weighted before and after drying to properly measure the boron concentration. In case of cell suspensions, the ratio between dry and fresh tissue was found to be lower than in case of tissue sections. When measuring track density in tissues it is therefore necessary to adjust the concentration using the measured dry to fresh weight ratio. It should be noted that the accuracy of the measurements relies on the possibility to avoid any source of contamination, in particular of boron, but also of all these substances that could capture neutrons and produce tracks in the detector. It is thus very important to work in a very clean environment and pay particular attention to the background measurement. This work provided calibration curves for liquid and solid biological samples, to be used as a reference when measuring tissues from animals and patients treated with boronated drugs for BNCT. Furthermore, the calibration for liquids will be used to measure the boron concentration in water solutions of new boronated compounds that are being developed in the frame of a new BNCT investigation branch dedicated to the treatment of Osteosarcoma [14].
Acknowledgements This work was partially supported by INFN (National Institute for Nuclear Physics), Italy, and by a CARIPLO funding program. The authors would like to acknowledge Drs. Giselle Saint Martin and Agustina M. Portu from CNEA (National Commission for Atomic Energy, Argentina) for their valuable help offered during the described work.
References Fig. 7. Tissue calibration curve; the angular coefficient is indicated as p1.
Fig. 8. Weight loss curve for calibration tissue samples.
[1] R. Barth, J. Coderre, M.G.H. Vicente, T. Blue, Boron neutron capture therapy of cancer: current status and future prospects, Clin. Cancer Res. 11 (2005) 3987– 4002. [2] K. Riley, O. Harling, An improved prompt gamma neutron activation analysis facility using a focused diffracted neutron beam, Nucl. Instrum. Meth. B 143 (1998) 414–421. [3] V.A. Nievaart, K. Appelman, F. Stecher-Rasmussen, W. Sauerwein, A. Wittig, R. Moss, Extension of the calibration curve for the PGRA facility in petten, Appl. Radiat. Isot 67 (7-8S) (2009) S362–S364. [4] T.U. Probst, Methods for boron analysis in boron neutron capture therapy (BNCT). A review, Fresenius, J. Anal. Chem. 364 (1999) 391–403. [5] S. Chandra, SIMS ion microscopy as a novel, practical tool for subcellular chemical imaging in cancer research, Appl. Surf. Sci. 203–204 (2003) 679–683. [6] D. Chiaraviglio, F. De Grazia, A. Zonta, S. Altieri, A. Braghieri, F. Fossati, P. Pedroni, T. Pinelli, A. Perotti Specchairello, G. Perlini, H. Rief, Evaluation of selective Boron absorption in liver tumors, Strahlenther. Onkol. 165 (2/3) (1989) 170–172. [7] A. Wittig, J. Michel, R.L. Moss, F. Stecher-Rasmussen, H.F. Arlinghaus, P. Bendel, P.L. Mauri, S. Altieri, R. Hilger, P.A. Salvadori, L. Menichetti, R. Zamenhof, W.A.G. Sauerwein, Boron analysis and boron imaging in biological materials for Boron Neutron Capture Therapy (BNCT), Crit. Rev. Oncol. Hematol. 68 (2008) 66–90.
56
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
[8] S. Altieri, S. Bortolussi, P. Bruschi, F. Fossati, K. Vittor, R. Nano, A. Facoetti, P. Chiari, J. Bakeine, A. Clerici, C. Ferrari, O. Salvucci, Boron absorption imaging in rat lung colon adenocarcinoma metastases, J. Phys. Conf. Series 41 (2006) 484–495. [9] S. Altieri, S. Bortolussi, P. Bruschi, P. Chiari, F. Fossati, S. Stella, U. Prati, L. Roveda, A. Zonta, C. Zonta, C. Ferrari, A. Clerici, R. Nano, T. Pinelli, Neutron autoradiography imaging of selective boron uptake in human metastatic tumors, Appl. Radiat. Isot. 66 (12) (2008) 1850–1855. [10] R.V. Griffith, L. Tommasino, Etch track detectors in radiation Dosimetry, in: K. Kase, B.B. Bjargard, F.H. Attix (Eds.), 1991. [11] C.H. Westcott, W.H. Walker, T.K. Alexander, Effective cross section and cadmium ratios for the neutron spectra of thermal reactors, Proceedings of
the International Conference on the Peaceful Uses of Atomic Energy, Geneva,1958, P/202 70, USA, New York, 1959. [12] Rene Brun, Fons Rademakers, ROOT – An object oriented data analysis framework, Proceedings AIHENP’96 Workshop, Lausanne, Sep. 1996, Nucl. Inst. Meth. Phys. Res. A 389 (1997) 81–86. See also
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
Set-up and calibration of a method to measure 10B concentration in biological samples by neutron autoradiography M.A. Gadan a,b, S. Bortolussi b,c,⇑, I. Postuma b, F. Ballarini b,c, P. Bruschi b, N. Protti b,c, D. Santoro b,c, S. Stella b,c, L. Cansolino d, A. Clerici d, C. Ferrari d, A. Zonta d, C. Zonta d, S. Altieri b,c a
National Commission for Atomic Energy (CNEA), Buenos Aires, Argentina Department of Nuclear and Theoretical Physics, University of Pavia, Pavia, Italy National Institute for Nuclear Physics (INFN), Section of Pavia, Pavia, Italy d Department of Experimental Surgery, University of Pavia, Pavia, Italy b c
a r t i c l e
i n f o
Article history: Received 19 October 2011 Received in revised form 24 November 2011 Available online 8 December 2011 Keywords: Boron neutron capture therapy Boron concentration measurements Neutron autoradiography CR-39
a b s t r a c t A selective uptake of boron in the tumor is the base of Boron Neutron Capture Therapy, which can destroy the tumor substantially sparing the normal tissue. In order to deliver a lethal dose to the tumor, keeping the dose absorbed by normal tissues below the tolerance level, it is mandatory to know the 10B concentration present in each kind of tissue at the moment of irradiation. This work presents the calibration procedure adopted for a boron concentration measurement method based on neutron autoradiography, where biological samples are deposited on sensitive films and irradiated in the thermal column of the TRIGA reactor (University of Pavia). The latent tracks produced in the film by the charged particles coming from the neutron capture in 10B are made visible by a proper etching, allowing the measurement of the track density. A calibration procedure with standard samples provides curves of track density as a function of boron concentration, to be used in the measurement of biological samples. In this paper, the bulk etch rate parameter and the calibration curves obtained for both liquid samples and biological tissues with known boron concentration are presented. A bulk etch rate value of (1.64 ± 0.02) lm/h and a linear dependence with etching time were found. The plots representing the track density versus the boron concentration in a range between 5 and 50 lg/g (ppm) are linear, with an angular coefficient of (1.614 ± 0.169)103 tracks/(lm2 ppm) for liquids and (1.598 ± 0.097)102 tracks/(lm2 ppm) for tissues. Ó 2011 Elsevier B.V. All rights reserved.
1. Introduction Boron Neutron Capture Therapy (BNCT) is an experimental form of radiotherapy consisting of two phases: administration of a 10B carrier able to concentrate this element in tumor more than in healthy tissues, and irradiation of the target with low energy neutrons [1]. The high LET radiation originating from neutron capture in 10B, irreversibly damages the cell in which the reaction takes place. If a sufficient boron concentration can be loaded in tumor cells and a low boron concentration is absorbed by normal tissues, the irradiation can lead to the selective destruction of essentially all the tumor cells present in the target. In order to assess a proper treatment plan, that is to deliver an effective irradiation with a substantial sparing of the healthy cells, it is necessary to know the boron concentration in each kind of irradiated tissue. In BNCT research, many methods are used to measure boron concentration in biological samples. One of the ⇑ Corresponding author. Address: Silva Bortolussi, Dip. Fisica Nucleare e Teorica, Università di Pavia, via Bassi 6, 27100 Pavia, Italia. Tel.: +39 0382 987635. E-mail address: [email protected] (S. Bortolussi). 0168-583X/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2011.11.043
most common ones is Prompt Gamma Neutron Activation Analysis (PGNAA) [2,3], a radio-analytic technique which exploits the 0.478 MeV c ray that is emitted from the excited 7Li in 93.9% of the capture reactions. With a facility dedicated to BNCT, it is usually possible to measure few ppms of 10B in relatively large samples (0.5 cm3) with irradiation times of some minutes. Of course, to measure low boron concentrations in small samples it is necessary to increase the measurement time. This technique is limited by the relatively high dimensions and the heterogeneity of the analyzed samples. In this case, the concentration that is obtained from the measurement procedure is an average value that cannot represent the uptake of the various cell fractions of the sample, such as viable tumor cells, healthy cells, necrosis, fibrosis and so on. The same problem is shared by other measurement methods such as ICP-AES (Inductively-Coupled Plasma-Atomic Emission Spectrometry) [4–7], which allows measuring concentrations in the range 0.2–70 ng/g. A more precise technique from the point of view of the spatial resolution is SIMS (Secondary Ion Mass Spectrometry) [5], which is a particular form of mass spectrometry. The sputtering of the sample surface with a beam of primary ions produces the emission
52
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
of ions from the sample. These ions are separated according to their mass, and their signals are transformed in an electronic map that constitutes a bi-dimensional image of the sample surface. The quantity of sample necessary for this technique is very small if the ion beam is focused in spots of 0.05–0.5 lm; boron concentrations that can be measured are in the interval 0.1–10 ppm. The Boron concentration measurement method used in Pavia for biological tissues is based on the spectroscopy of the charged particles coming from the reaction 10B(n,a)7Li, when irradiating thin tissue sections in the thermal column of TRIGA reactor [6,7]. The sections are laid on mylar supports and irradiated in a vacuum chamber facing a Si detector. This method is not suitable for liquid samples, like blood and urine, because they cannot stay under vacuum. To solve this problem, a method based on neutron autoradiography is proposed. The autoradiography technique is based on the interaction of high-LET particles crossing a material and producing damages in its molecular structure along the particle tracks. Since in conductors and semiconductor materials a recombination process occurs only dielectric materials show regions with an irreversible damage. These regions are called latent tracks. By exposing the material to a proper etchant solution, chemical reactions occur everywhere in the material, but more intensively in the latent track regions. This effect of the etching process produces track pits on the surface of the material, which are visible with an optical microscope. Because of their properties, dielectric materials such as plastics and glasses are used in the autoradiography technique, such as sensitive supports called SSNTDs (Solid-State Nuclear Track Detectors) [10]. In the case of boron concentration measurements in biological samples, a small amount of sample is deposited on a SSNTD, for the irradiation in the neutron field. The SSNTD is then etched with a proper solution to visualize the damage produced by high-LET particles during the sample–neutron interaction. Finally, the images taken with an optical microscope can be analyzed with a suitable software to perform different kinds of analysis. Depending on the neutron fluence, the autoradiography technique can be employed to make quantitative or qualitative analysis. On one hand, at a high neutron fluence, a higher density of high-LET particles coming from 10B(n,a)7Li reaction is produced, and with a proper etching it is possible to obtain images for macroscopic boron bio-distribution qualitative analysis. A method to study the boron bio-distribution in biological samples by autoradiography imaging was already assessed in Pavia [8,9]. The irradiation of the samples and the etching parameters were optimized to obtain images of 10B distribution, which were compared to the histological preparation of a contiguous tissue section. In this way it is possible to verify if the areas with higher track density correspond to the tumor, and thus if the tumor cells are loaded with higher boron concentrations. On the other hand, irradiating with a low neutron fluence, the density of high-LET particles produced is lower, thus causing less tracks in the detector. A proper etching of the SSNTD allows visualizing the tracks separately, and counting them. Images are collected with an optical microscope and are analyzed with an adequate software to determine the track density, which is used to calculate the boron concentration in the sample through a proper calibration curve. The main aim of the present work was to measure the calibration curve for liquid and solid biological samples. The paper describes the experimental set-ups, the preparation of the calibration samples, and the results of the measurements.
2. Materials and methods The sensitive films used for neutron autoradiography were rectangular polyallyldiglycol carbonate (PADC) CR-39 film
detectors from Intercast Europe manufacturer, with a 75 25 mm2 area and a 1 mm thickness. Before starting with boron concentration measurements, the etching process in CR-39 films was characterized by the determination of the bulk etch rate parameter, which is useful to compare the results of autoradiographies obtained under different etching conditions. Another important parameter is the neutron flux at which the samples were exposed, in order to set the fluence received by the samples with good accuracy. To characterize the calibration setup, the neutron flux was measured in the facilities used for liquid and solid biological sample irradiation. Calibration curves were assessed both for liquid and for solid biological samples in a range between 5 and 50 ppm. 2.1. Bulk etch rate measurement The formation of a track depends on the competitive effect of two parameters: the track etch rate (VT) and the bulk etch rate (VB). The track etch rate is the velocity at which the etching process progresses along the latent track, whereas the bulk etch rate is the etching velocity at undamaged regions of the detector. Considering a particle with normal incidence on the detector, and a constant track etch rate (VT) and bulk etch rate (VB), the diameter of the track (D) at an etching time (t) can be expressed as follows [10]:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi VT VB D ¼ 2 VB t VT þ VB
ð1Þ
Eq. (1) shows the competitive effect between VT and VB in the growth of the track diameter. Since for heavy ions and fission fragments VT VB, the removed layer, and thus VB, can be related to the track opening diameter by the relation [10]:
DF VB t 2
ð2Þ
where DF is fission fragment track diameter, VB is bulk etch rate and t is etching time. For VB measurement, the CR-39 films were irradiated with a 252 Cf source at a short distance and etched for the following times: 30, 45, 60, 70 and 80 min, in a 6.25 N NaOH solution at 70 °C. The etching solution was prepared mixing 99.0% purity NaOH with pure water. After etching, CR-39 films were washed out with cold water in order to remove the etchant solution from the detector surface. The experimental set up for the image acquisition is composed by a stereomicroscope (Leica MZ16A) connected to a lamp (Leica CLS150X) and a joystick (Prior Optiscan II). The microscope has an integrated camera connected to a PC and the software LAS V3.7 was employed to acquire the pictures. The analysis of fission fragment tracks was performed with the ImagePro Plus Analizer 6.3 code. 2.2. The irradiation facilities and the neutron flux characterization Liquid samples were irradiated in the facility shown in Fig. 1, which allowed pouring 5 ml of solution in sealed holes below which CR-39 films were positioned. This device hosted four CR-39 films, and for each film four holes for liquid samples were available. In this way it was possible to obtain films in which up to four different concentrations were simultaneously present. The neutron flux was measured by means of copper foil activation using the Wetscott method [11]. Copper foils were positioned inside four holes, one in each CR-39, filled with water and irradiated in the same position as for the real measurements (see Section 2.3).
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
53
Fig. 1. Liquid samples irradiation facility. It is visible the inferior plexiglass bar with 4 housings for CR-39 slides and the superior one with the sealed holes where the boronated solutions were poured.
Fig. 2. Picture of the etched CR-39, for a liquid sample with 50 ppm of irradiated with a fluence of (5.18 ± 0.21) 1010 n/cm2.
After irradiation, the foils were taken off the facility and their activity was measured by a germanium detector. Solid biological samples were irradiated in a facility made up of two plexiglass slides positioned vertically at the end of the thermal column. This facility allows the irradiation of up to thirty CR-39 films positioned at different heights in the same plane. The neutron flux determination was carried out using the same method as for liquid irradiation facility, positioning copper foils in representative points of the CR-39 irradiation facility.
parameters were set to obtain the highest contrast between the background and the interesting points of the picture. A shadow correction (corresponding to a background subtraction) was applied to remove the inhomogeneous light field, which would alter the quality of the pictures. Once the images were acquired, the code Image-Pro 6.3 was used to classify the charged particle tracks, visualized as dark objects on a uniform gray background (Fig. 2). The analysis process required the binarization of the images, using the gray scale histogram. A built-in code was then employed to calculate some userdefined parameters, such as the minimum diameter of the tracks and their area. These two parameters where used to select the interesting points to be counted. The software employed for the analysis was a home-made script written in the data analysis framework ROOT [12]. The distribution of the track area was used to set a threshold corresponding to the objects of 1 pixel, which are not tracks but artifacts that the binarization program could not delete. Another threshold was set for too large objects, which could be multiple tracks, or artifacts, or tracks not produced by a particles. These events are quite unusual, but they must be excluded from the counts. The minimum diameter distribution was used to define the range of diameters where the objects to be counted lay. To this aim, all the photos taken for each concentration value were used to plot a histogram of the minimum diameters, and this distribution was plotted with a Gaussian curve (Fig. 3). The interval [ld 3rd, ld + 3rd] was chosen for each concentration value as the range where the tracks must be counted. In each picture, the tracks within this interval were then counted and plotted as a histogram and fitted with a Gaussian curve (Fig. 4). The mean value lc of this fit was divided by the area of the field and used as mean track density for the considered concentration; 1r was used as the error associated to this point. Finally, the points obtained for all the concentration values were plotted as a function of boron concentration, and the linear fit was computed.
2.3. Preparation and irradiation of the calibration samples Liquid standards consisting of water solutions of boric acid at boron concentrations of 5, 10, 25 and 50 ppm were poured on CR-39 films, using the device constructed on purpose for the irradiation of liquid samples. Liquid samples were irradiated at the end of the thermal column of the TRIGA Mark II reactor, operating at 10 kW for 30 min, receiving a neutron fluence of (5.18 ± 0.21) 1010 n/cm2. After irradiation, the CR-39 films were chemically etched for 60 min with the same etching parameters used for the fission fragment measurement. Samples obtained mixing cell suspensions of rat healthy liver with BPA-HCl, at different boron concentrations, were used for tissue calibration. Briefly, cell suspensions were obtained by mechanically treating the liver tissue with a potter homogenizer. The tissue inserted into the tube was pressed until its complete disaggregation. Then the obtained cell suspensions were divided into five aliquots, enriched by centrifugation and weighted. A boronophenylalanine–HCl (BPA–HCl) solution at 500 ppm of 10B was added to each cell fraction in order to obtain the reference boron concentrations of 5, 10, 25 and 50 ppm. The concentration values were then recalculated on the basis of the true weights of each prepared sample; the error associated to the concentration values is of the order of 0.5%. Finally, the samples were frozen in small cylindrical rods at 80 °C. Slices of 60 lm were obtained using a Leica cryostat and were deposited on CR-39 films, which were then irradiated at the end of the thermal column with the reactor operating at 5 kW for 30 min, receiving a neutron fluence of (4.91 ± 0.28) 1010 n/cm2. After irradiation, the films were chemically etched at the described etching conditions for 45 min. 2.4. Track density analysis After the etching process, a number of images from the irradiated samples were taken with the stereomicroscope. In order to measure the track density with good accuracy, acquisition
10
B,
2.5. The dry to fresh mass ratio of the samples When using the calibration curve obtained for solid samples, an important parameter must be taken into account: the amount of water loss occurring after the preparation of the tissue sections. Tissues with the same composition as the calibration samples could lose water in different percentages. Therefore, when calculating boron concentration in a sample by comparison with the calibration curve, the results must be renormalized by the ratio between the water loss in the calibration samples and in the tissue
54
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
Fig. 3. Example of histogram representing the minimum diameter of the tracks, obtained from all the pictures taken for a single concentration value.
Fig. 5. Bulk etch rate curve.
achieved in the track diameter measurement, the relative error decreases with the etching time, and it is maximum (13%) for an etching time of 30 min. 3.2. Neutron flux measurements in the facilities The thermal neutron flux measured in the liquid irradiation facility with the reactor operating at 250 kW was (7.20 ± 0.29)108 n/cm2s, thus in the liquid irradiation position there was a flux of (2.88 ± 0.12)107 n/cm2s with the reactor operating at 10 kW. The thermal neutron flux measured in the solid sample irradiation facility with the reactor operating at 250 kW was (1.37 ± 0.08)109 n/cm2s and the flux in the plexiglass holding at a power of 5 kW was (2.74 ± 0.16)107 n/cm2s. 3.3. Calibration for liquid samples Fig. 4. Example of a histogram representing the number of tracks/picture obtained from all the pictures taken for a single concentration value. The mean value of the Gaussian, divided by the area of the picture, was used in the calibration curve; the sigma of this distribution is the error associated to the point.
sample. In order to measure the ratio between the weight of a dry calibration sample and the fresh one, a number of 60 lm sections were cut and weighted by a high sensitivity digital scale (up to 1 lg). Since loss of weight occurs very quickly for thin sections, the scale was connected to a Personal Computer and the weight was acquired every second by means of a home-made acquisition software. When the section was completely dried, that is the plot of the weight versus time was constant, the dry to fresh weight ratio was calculated. The process was repeated several times, and the ratio was obtained as an average of all the performed measurements.
The calibration curve for liquid samples is shown in Fig. 6. The fit of the plot is linear, with an angular coefficient of (1.614 ± 0.169)103 tracks/(lm2 ppm). The v2/ndf of the fit is 0.9/3. Taking into account a VB of 1.63 lm/h and an etching time of 60 min, the removed layer was (1.64 ± 0.02) lm. The errors associated with the points at low concentrations are quite high, depending on the low track density in the pictures. This could be avoided irradiating with higher neutron fluence. Nevertheless, this experiment was conceived to measure the track density in a range between 5 and
3. Results 3.1. Bulk etch rate The thickness of the removed layer as a function of the etching time according to (2) is shown in Fig. 5. The track diameter increases with the etching time for the whole considered time range, and a linear response with a constant bulk etch rate of VB = (1.64 ± 0.02) lm/h was observed. This value is in good agreement with those reported in the literature for fission fragments and the same etching parameters [13]. Thanks to the precision
Fig. 6. Calibration curve for liquid samples; the angular coefficient is indicated as p1.
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
50 ppms, and the fluence was chosen in a way that the tracks remained separated also in the samples with higher boron concentration. When samples with low concentration are measured, it would be always possible to make another calibration for a small range of boron concentrations, irradiating with a higher neutron fluence and thus lowering the errors. 3.4. Calibration for tissue samples The calibration curve for tissue samples is shown in Fig. 7. Again, the fit is linear, with an angular coefficient of (1.598 ± 0.097)102 tracks/(lm2 ppm). The v2/ndf of the fit is 2.6/3. For an etching time of 45 min the removed layer was (1.23 ± 0.01) lm. 3.5. The dry to fresh mass ratio of the samples The dry to fresh weight ratio measured for the frozen cell suspension was (0.15 ± 0.01). Fig. 8 shows the weight loss by a tissue calibration sample. Experimental measurements showed that, after 1500–1800 s, the variation in the weight loss was less than 1%, under our laboratory conditions. The rise in the first part of the curve is due to the sample position in the scale. After this steep rise, the tissue sample weight decreases down to a limit value. The ratio between this limit value and the fresh weight is the parameter to be used in the concentration calculation. 4. Discussion and conclusions Calibration curves obtained both for liquid and for solid samples showed a very good linearity in the range between 5 and 50 ppm, allowing boron concentration measurements in biological samples,
55
both liquid and solid, for BNCT purposes using neutron autoradiography in the described conditions. It is very important to note that a proper calibration is needed for each kind of sample that must be measured. In fact, the stopping power of the charged particles depends on the material that they cross. For this reason, the calibration for solid biological tissues was performed using liver tissue reduced to a cell suspension. In this way, when tissue samples are cut, deposited on CR-39 and irradiated, the obtained track density can be related to the calibration curve, because the stopping power of a particles is the same. Therefore, it is important to use tissue calibration when measuring tissues, water solution calibrations when measuring liquids, and so on. Furthermore, it must be taken into account that the sectioned tissues dry very rapidly, modifying their water content. Our experience shows that the biological variability of this weight change is high and each type of tissue must be weighted before and after drying to properly measure the boron concentration. In case of cell suspensions, the ratio between dry and fresh tissue was found to be lower than in case of tissue sections. When measuring track density in tissues it is therefore necessary to adjust the concentration using the measured dry to fresh weight ratio. It should be noted that the accuracy of the measurements relies on the possibility to avoid any source of contamination, in particular of boron, but also of all these substances that could capture neutrons and produce tracks in the detector. It is thus very important to work in a very clean environment and pay particular attention to the background measurement. This work provided calibration curves for liquid and solid biological samples, to be used as a reference when measuring tissues from animals and patients treated with boronated drugs for BNCT. Furthermore, the calibration for liquids will be used to measure the boron concentration in water solutions of new boronated compounds that are being developed in the frame of a new BNCT investigation branch dedicated to the treatment of Osteosarcoma [14].
Acknowledgements This work was partially supported by INFN (National Institute for Nuclear Physics), Italy, and by a CARIPLO funding program. The authors would like to acknowledge Drs. Giselle Saint Martin and Agustina M. Portu from CNEA (National Commission for Atomic Energy, Argentina) for their valuable help offered during the described work.
References Fig. 7. Tissue calibration curve; the angular coefficient is indicated as p1.
Fig. 8. Weight loss curve for calibration tissue samples.
[1] R. Barth, J. Coderre, M.G.H. Vicente, T. Blue, Boron neutron capture therapy of cancer: current status and future prospects, Clin. Cancer Res. 11 (2005) 3987– 4002. [2] K. Riley, O. Harling, An improved prompt gamma neutron activation analysis facility using a focused diffracted neutron beam, Nucl. Instrum. Meth. B 143 (1998) 414–421. [3] V.A. Nievaart, K. Appelman, F. Stecher-Rasmussen, W. Sauerwein, A. Wittig, R. Moss, Extension of the calibration curve for the PGRA facility in petten, Appl. Radiat. Isot 67 (7-8S) (2009) S362–S364. [4] T.U. Probst, Methods for boron analysis in boron neutron capture therapy (BNCT). A review, Fresenius, J. Anal. Chem. 364 (1999) 391–403. [5] S. Chandra, SIMS ion microscopy as a novel, practical tool for subcellular chemical imaging in cancer research, Appl. Surf. Sci. 203–204 (2003) 679–683. [6] D. Chiaraviglio, F. De Grazia, A. Zonta, S. Altieri, A. Braghieri, F. Fossati, P. Pedroni, T. Pinelli, A. Perotti Specchairello, G. Perlini, H. Rief, Evaluation of selective Boron absorption in liver tumors, Strahlenther. Onkol. 165 (2/3) (1989) 170–172. [7] A. Wittig, J. Michel, R.L. Moss, F. Stecher-Rasmussen, H.F. Arlinghaus, P. Bendel, P.L. Mauri, S. Altieri, R. Hilger, P.A. Salvadori, L. Menichetti, R. Zamenhof, W.A.G. Sauerwein, Boron analysis and boron imaging in biological materials for Boron Neutron Capture Therapy (BNCT), Crit. Rev. Oncol. Hematol. 68 (2008) 66–90.
56
M.A. Gadan et al. / Nuclear Instruments and Methods in Physics Research B 274 (2012) 51–56
[8] S. Altieri, S. Bortolussi, P. Bruschi, F. Fossati, K. Vittor, R. Nano, A. Facoetti, P. Chiari, J. Bakeine, A. Clerici, C. Ferrari, O. Salvucci, Boron absorption imaging in rat lung colon adenocarcinoma metastases, J. Phys. Conf. Series 41 (2006) 484–495. [9] S. Altieri, S. Bortolussi, P. Bruschi, P. Chiari, F. Fossati, S. Stella, U. Prati, L. Roveda, A. Zonta, C. Zonta, C. Ferrari, A. Clerici, R. Nano, T. Pinelli, Neutron autoradiography imaging of selective boron uptake in human metastatic tumors, Appl. Radiat. Isot. 66 (12) (2008) 1850–1855. [10] R.V. Griffith, L. Tommasino, Etch track detectors in radiation Dosimetry, in: K. Kase, B.B. Bjargard, F.H. Attix (Eds.), 1991. [11] C.H. Westcott, W.H. Walker, T.K. Alexander, Effective cross section and cadmium ratios for the neutron spectra of thermal reactors, Proceedings of
the International Conference on the Peaceful Uses of Atomic Energy, Geneva,1958, P/202 70, USA, New York, 1959. [12] Rene Brun, Fons Rademakers, ROOT – An object oriented data analysis framework, Proceedings AIHENP’96 Workshop, Lausanne, Sep. 1996, Nucl. Inst. Meth. Phys. Res. A 389 (1997) 81–86. See also