If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Boron Neutron Capture Therapy (BNCT) is a binary radiotherapy method developed to treat patients with certain malignant tumours. To date, over 300 treatments have been carried out at the Finnish BNCT facility in various on-going and past clinical trials. In this technical review, we discuss our research work in the field of medical physics to form the groundwork for the Finnish BNCT patient treatments, as well as the possibilities to further develop and optimize the method in the future. Accordingly, the following aspects are described: neutron sources, beam dosimetry, treatment planning, boron imaging and determination, and finally the possibilities to detect the efficacy and effects of BNCT on patients.
Boron neutron capture therapy of brain tumors: enhanced survival and cure following blood-brain barrier disruption and intracarotid injection of sodium borocaptate and boronophenylalanine.
Boron neutron capture therapy (BNCT) followed by intensity modulated chemoradiotherapy as primary treatment of large head and neck cancer with intracranial involvement.
In BNCT, the boron (10B) carrier compound is usually administered by an intravenous infusion. In the absence of neutrons, the boron compound is a non-toxic and non-radioactive agent that accumulates into cancer cells. After a period of time, boron is optimally located in the tumour cells, while the healthy tissue has a lower boron concentration. The tumour site is irradiated with neutrons, which thermalize in tissue and interact with the 10B nuclei. As a result, high linear energy transfer (LET) alpha and lithium particles are produced, destroying the surrounding cell. In the Finnish trials, intravenously administered l-boronophenylalanine–fructose (l-BPA-F) has been used as the boron carrier, and the patient is irradiated with a beam of epithermal neutrons from the reactor without craniotomy or performing other surgical procedures.
The prerequisite for successful BNCT is to concentrate a sufficient amount of 10B in the tumour cells and to irradiate the patient with neutrons at a high enough intensity and within a suitable energy range, enabling the boron neutron capture reaction to occur. Accordingly, it is important to validate and optimize the behaviour of the boron biodistribution, the energy distribution of the neutron field and the set positioning (Fig. 1). The boron concentration ratio between tumour and healthy cells is crucial, for example, for 20–30 ppm boron concentration 3–4 concentration ratio is needed for the 109 neutrons/cm2/s collimated neutron flux. A great deal of effort has been invested in treatment planning to maximize the benefit to the patient. The Finnish FiR 1 epithermal neutron beam [
Seppälä T. FiR 1 epithermal neutron beam model and dose calculation for treatment planning in neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2002.
]. A custom-made patient positioning system including a treatment table, beam aperture simulator and positioning lasers has been developed and constructed [
], and the effect of positioning uncertainty on the doses has been studied with Monte Carlo simulations. An image registration protocol for BNCT has been created [
Figure 1The fundamental dosimetric process of BNCT occurs at the cellular level due to the short path lengths of alpha and lithium recoil nuclei produced in the boron neutron capture reaction. Macroscopic distribution of the boron carrier in tissues is temporally optimized for the irradiation, the tumor having maximal boron uptake compared to the healthy tissue.
According to International Commission of Radiation Units and Measurements (ICRU), the uncertainty of the dose to the patient in external radiotherapy should not exceed 5%, the recommendation from literature being below 3% [
]. These facts set high objectives for reliability and accuracy in the patient positioning in addition to the boron level definition and beam dosimetry in BNCT. The majority of European research groups involved in the development of BNCT have prepared the report “Recommendations for the Dosimetry of Boron Neutron Capture Therapy (BNCT)” published by the Nuclear Research & consultancy Group (NRG) [
Järvinen H. Voorbraak W.P. Recommendations for the dosimetry of boron neutron capture therapy (BNCT). The Nuclear Research and consultancy Group NRG,
2003
]. The report covers the guidelines for the basic characterization and dosimetry of the epithermal (and thermal) neutron beams used in BNCT. However, this report includes no guidelines for determining boron biodistribution or dose planning calculations. In the future, microdosimetric approaches, quantitative in vivo imaging before, during and after the treatment, as well as taking full advantage of computational methods are expected to change the situation [
]. There are also some prospects to find better carrier compounds and optimize the delivery strategies to bring boron efficiently and selectively to the target cells [
Boron neutron capture therapy of brain tumors: enhanced survival and cure following blood-brain barrier disruption and intracarotid injection of sodium borocaptate and boronophenylalanine.
In this review, the following aspects will be described: neutron sources, beam dosimetry, treatment planning, boron imaging and determination, and finally the possibilities to detect the BNCT effects on patient, the emphasis being on the research and technical development carried out in Finland during the last 20 years.
Neutron sources
At our TRIGA (Training, Research, Isotopes, General Atomics) Mark II type FiR 1 reactor-based BNCT facility, fission neutrons are slowed down to the epithermal energy range using FLUENTAL™ (69 w-% of AlF3, 30 w-% of metallic Al, 1 w-% 7LiF) moderator material developed at VTT [
]. The moderated neutrons are collimated and gamma-shielded with bismuth into the high intensity forward directed (current to fluence ratio 0.77) epithermal neutron beam with low gamma, thermal neutron and fast neutron contamination. The beam characteristics and intensity have been confirmed with the measurements performed by the Finnish dosimetry team and by visiting teams from Idaho National Laboratory, Idaho USA [
]. The advantage of using a TRIGA reactor for BNCT is its stability and reliability in addition to the high neutron intensity and low background radiation of the treatment beam [
]. None of the patient treatments has been cancelled or postponed due to reactor related problems during the history of BNCT in Finland.
One of the drawbacks in using a nuclear reactor as a BNCT neutron source is obviously its location in non-hospital environment, even though in Finland the reactor is located only 15 min away from the nearest hospital unit. Also, since the beam position is stationary, the patient needs to be directed and rotated in the treatment fields. To address this issue, for nearly three decades, the development of Accelerator Based Neutron Sources (ABNS) has been of interest, because ABNS could be safely installed in hospitals [
Boron neutron capture therapy for the treatment of glioblastomas and extracranial tumours: as effective, more effective or less effective than photon irradiation?.
]. Firstly, a high power accelerator for producing a high-current particle beam is required. Secondly, an appropriate target for producing neutrons needs to be developed with an efficient heat removal system. Thirdly, in order to reduce the initial neutron energy down to the optimal range for BNCT, a beam shaping assembly (BSA) must be designed.
The optimal neutron beam energy for treating deep-seated tumours with BNCT is from 4 eV to 40 keV [
]. A common problem for neutron sources used for BNCT is that it is not possible to produce a high neutron flux directly at the optimal epithermal energy range. In addition, considerable moderation of the source neutrons is always required. A neutron beam can be moderated, for example, by placing a tissue equivalent material (bolus) on the patient's skin to treat more superficial tumours [
]. During the moderation, neutron intensity is reduced and consequently, the proton accelerator for ABNS for BNCT needs to be operated at much higher power than for example a proton therapy accelerator. The 7Li(p,n)7Be neutron source with 2.5 MeV protons at 20 mA current for BNCT requires 400 times more powerful accelerator than used at a typical proton therapy unit with up to 250 MeV protons operated typically at 500 nA.
Neutron producing reactions can be induced by accelerated protons, deuterons or tritons targeting 7Li, 9Be, 13C, 12C, 2H or 3H nuclei. Probably the most studied ABNS application is the 7Li(p,n)7Be reaction at around 2.5 MeV proton energy, because a sufficiently low accelerator current (10 mA) is required for producing a high intensity of reasonably low energetic (<1 MeV) neutrons. Such BNCT beams with functional lithium targets and beam shaping assemblies have been developed or are under development at the University of Birmingham in the UK [
] produces neutrons through the 9Be(p,n)9B reaction utilizing a 30 MeV proton beam at 1 mA current. Such a powerful neutron source also results in strong activation of the materials within the device. In addition, there are difficulties in reducing the unwanted fast neutron and photon components from the clinical neutron beam [
]. Furthermore, there is a plan to apply a 5 MeV proton beam line for thermal neutron BNCT at Legnaro Laboratories of National Institute of Nuclear Physic (INFN) in Italy [
]. The neutrons are intended to be produced by 9Be(p,xn) reactions, where a neutron yield of 1014 s−1 of <3.2 MeV neutrons is expected. ABNS with a beryllium target for a high flux (>109 neutrons/cm2/s) thermal neutron beam for skin melanoma BNCT have been developed in the project [
To date, all the clinical BNCT trials have been carried out using reactor-based neutrons due to the high thermal or epithermal neutron flux of order 109 neutrons/cm2/s [
]. However, while we do not have an ABNS development project in Finland, we recognize that there is no doubt that realizing ABNS is one of the determining factors in how fast BNCT will be established as routine clinical treatment.
Beam dosimetry
To assure the reliability and consistency of dosimetry and the metrological traceability of calibrations, the dosimetry procedures in radiation therapy follow international guidelines [
]. For uncommon and specific type of radiotherapy modalities, such as BNCT, no uniform international dosimetry guidance exists so far. The work towards coherent BNCT dosimetry procedures was started in 1998, when a European project for a Code of Practice of epithermal neutron beams in BNCT was launched [
Järvinen H. Voorbraak W.P. Recommendations for the dosimetry of boron neutron capture therapy (BNCT). The Nuclear Research and consultancy Group NRG,
2003
]. The absorbed dose must have a metrologically traceable link to the international dosimetry standards also in BNCT (Fig. 2). In dosimetry of epithermal neutron beams, a reasonable accuracy in the determination of photon and neutron doses has been achieved using ionization chambers [
]. The dual ionization chambers are recommended and often used to determine photon and (fast) neutron dose in a mixed photon and epithermal neutron beam [
Järvinen H. Voorbraak W.P. Recommendations for the dosimetry of boron neutron capture therapy (BNCT). The Nuclear Research and consultancy Group NRG,
2003
], which are not directly applicable to the epithermal neutron beams used in BNCT.
Figure 2Relation and use of different dosimetric methods. In BNCT the physical dose must have a metrologically traceable link to the international dosimetry standard.
During the establishment of systematic dosimetry procedures in the dosimetry of radiation therapy, the importance of quality assurance (QA) has been emphasized [
]. The American Association of Physicists in Medicine (AAPM) published recommendations for a comprehensive QA for radiation therapy in 1994, and in 1997, the European Society for Therapeutic Radiology and Oncology (ESTRO) presented a frame for a quality system based on the guidelines of the International Standardization Organization (ISO) [
Quality assurance in radiotherapy. European society for therapeutic radiology and oncology advisory report to the commission of the european union for the ‘Europe against cancer programme’.
Järvinen H. Voorbraak W.P. Recommendations for the dosimetry of boron neutron capture therapy (BNCT). The Nuclear Research and consultancy Group NRG,
2003
Uusi-Simola J. Finnish dosimetric practice for epithermal neutron beam dosimetry in boron neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2009.
In the Finnish BNCT program, we use an Exradin T2 ionization chamber made of A-150 tissue equivalent plastic and flushed with tissue equivalent gas (referred to as the TE(TE) chamber) to measure total neutron and photon dose. An Exradin M2 magnesium ionization chamber flushed with argon gas (Mg(Ar) chamber) is used to measure the photon dose. The ionization chambers are individually calibrated for absorbed dose to water in 60Co gamma radiation beam [
]. For the TE(TE) chamber the relative sensitivity for neutrons is determined according to general Bragg-Gray cavity theory following the ICRU formalism [
]. Thus the photon dose determination does not require information of the neutron spectrum. The neutron sensitivity of photon chambers has been investigated in other studies, where also non-zero values have been acquired [
]. These results demonstrate that the assessment of relevant dosimetry factors in a mixed field is potentially sensitive to specific measurement set-up, and thus careful considerations are required when analysing the acquired experimental data. The absorbed doses from the neutrons cannot be measured directly. Measurements with activation detectors and semiconductor diodes produce reaction rates and neutron fluence rates which are converted into doses by application of the kerma factors [
Microdosimeters based on a solid state semiconductor detector or a proportional counter can be applied in BNCT dosimetry to measure the total neutron and photon dose [
Study of a solid state microdosemeter based on a monolithic silicon telescope: irradiations with low-energy neutrons and direct comparison with a cylindrical TEPC.
]. However, the use of the commercially available proportional counter (1/2" TEPC, Far West Technology) is limited to measurements at low beam intensity [
]. Silicon diode based microdosimeters have a high potential for use in BNCT, since all dose components in the mixed field can be measured on-line with a single small detector [
]. However, the optimal characteristics of the semiconductor detectors for BNCT are still under development. The construction of the detector includes a layer of suitable converter material attached to the surface of a p-n junction diode where the converted charged particles can be detected and measured [
Aschan C. Applicability of thermoluminescent dosimeters in X-Ray organ dose determination and in the dosimetry of systemic and boron neutron capture radiotherapy. PhD thesis. University of Helsinki, Helsinki; 1999.
]. In the Finnish BNCT project we have studied two types of polymer gel dosimeters and applied gel dosimetry to measure the relative total dose distribution [
]. Superheated drop detectors, fission counter, BF3 counters, boron lined proportional counters and 3He proportional counters can be used for the detection of neutrons [
Järvinen H. Voorbraak W.P. Recommendations for the dosimetry of boron neutron capture therapy (BNCT). The Nuclear Research and consultancy Group NRG,
2003
The main potential and inspiration of BNCT is that the therapeutic effects emerge at the cellular level. BNCT dose calculation requires careful calculation of nuclear interactions of neutrons with the tissue elements and thus, elemental composition of tissues within patient geometry needs to be modelled. In the Finnish BNCT practice, the patient models are created semiautomatically based on information from the computed tomography (CT) or manually based on magnetic resonance (MR) images. The clinical target volume (CTV) is determined based on contrast-enhanced T1-weighted MR images, while also T2-weighted MR images and positron emission tomography (PET) images are frequently utilized [
], the CT image set of the first treatment fraction defines the baseline for the images coordinate system. In the case of two head and neck BNCT fractions, dose planning CT images are taken for both fractions. When the new CT studies are registered with the first reference CT study, an intra-modality registration is carried out. Inter-modality image registration is performed between different image modalities (MRI/CT/PET). In the latter cases, the image similarity metrics based on the direct comparison of gray levels (like normalized correlation) cannot be utilized in the optimization of the registration parameters, because the intensities in the different imaging modalities do not correlate linearly.
Instead, we apply Mattes implementation of mutual information metrics [
] and three rotation and three translation parameters (3D rigid transform). In order to decrease the possibility of finding a local maximum of mutual information metric instead of the global maximum, a multi-resolution approach is applied. In the multi-resolution approach, the registration is performed first at a coarse scale. The registration parameters from the coarse scale are used as start values of the optimization at a finer scale. The optimization of the metric is carried out using physical coordinates instead of voxel coordinates. The pixel spacing, image position and direction cosines which couple the voxel coordinates to physical coordinates are extracted from DICOM (The Digital Imaging and Communications in Medicine) image files [
]. Hence the estimation of scale parameters (relation of voxel sizes in different studies) is not necessary, which reduces the dimensionality of search space of the optimizer. Multiresolution approach has been widely proven to provide faster, more robust and more accurate computation of the registration parameters [
]. Successive registrations are needed, for example, when images taken at one point of time are first registered with each other, and then, possibly months or years later, these registered images are re-registered with the new images – either for the purposes of new treatment planning or registration aided detection of change [
The use of 3D rigid transform for registration from different modalities requires careful positioning of the patient for imaging. In the case of head-and-neck cancer patients, it is crucial to avoid different bending of neck. Fixation masks, tattooed points and external markers should be used to assist the patient positioning [
]. It would be difficult to automate elastic registration to work robustly, and to restrict the corrected deformations only to the differences due to patient positioning. Additionally, rigid transformations are faster to optimize. Therefore our protocol is presently based on the careful patient positioning and on the use of rigid 3D transform.
The registration results are verified using multiplanar reconstructions and comparing anatomical features using visual interactive tools based on the Visualization Toolkit [
]. Final verification of the registration is done in the treatment planning software with cubically resampled data. The verification is done by comparing the segmented structures (skin, bones, brain, medulla, air cavities) of one CT image set on the other CT image set.
] and the Visualization Toolkit (VTK). As a future work, it is important to study the impact of the inaccuracies in the image registration on the dose estimation accuracy. If seen necessary, the registration errors can possibly be decreased by extending the registration environment with deformable (non-rigid) transformations or improving patient fixation systems. The approach based on the in-house developed software and open source toolkits (ITK and VTK) enables further development and optimization of the registration framework.
Dose calculation
So far, only Monte Carlo (MC) based treatment planning systems (TPS) have been applied in clinical BNCT. Four MC -based TPSs have been developed for BNCT use: the NCT_Plan developed at Harvard and MIT [
], and the JAEA computational dosimetry system (JCDS) developed at Japan Atomic Energy Agency (JAEA). The MC methods are often very time consuming, which has led to interest in the application of faster deterministic methods [
The dosimetric calculations are often performed with the extensively benchmarked and validated multipurpose software, Monte Carlo N-Particle Code (MCNP) [
] or respectively, the 2- or 3-dimensional discrete ordinate method based DORT (two-dimensional discrete ordinates (deterministic) transport) and TORT (three-dimensional discrete ordinates (deterministic) transport) codes [
]. The SERA system uses its own MC code, seraMC, for neutron and photon transport calculations. In the Finnish BNCT project, the MCNP or the DORT codes are applied for the simple geometry dosimetric calculations, and the SERA code for the clinical treatment planning [
Kotiluoto P. Adaptive tree multigrids and simplified spherical harmonics approximation in deterministic neutral and charged particle transport. PhD thesis. VTT, Espoo; 2007.
Seppälä T. FiR 1 epithermal neutron beam model and dose calculation for treatment planning in neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2002.
]. The beam model includes neutron and photon transport from the reactor core through the moderator and collimator structures towards beam aperture. The neutron and photon energy and angular distribution obtained with the DORT are converted to a beam model for MCNP and SERA by averaging the forward current for 47 neutron energy groups and 20 photon energy groups. For each energy group, separate angular distributions by 10 cosine cut points in the forward direction have been defined [
Kotiluoto P. Adaptive tree multigrids and simplified spherical harmonics approximation in deterministic neutral and charged particle transport. PhD thesis. VTT, Espoo; 2007.
]. Neutron models of TPS have been normalized to measurements using diluted AuAl (1w-%, Au) foils in the cylindrical PMMA phantom at the depth of 2.0 cm [
Seppälä T. FiR 1 epithermal neutron beam model and dose calculation for treatment planning in neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2002.
]. The normalization factor of the computational beam model is the ratio of independently measured and computed 197Au(n,ã) activation reaction rates. The neutron energy distribution of FiR 1 beam models in phantoms has been verified by comparing the measured and computed activation reaction rates of 197Au(n,ã) and 55Mn(n,ã), which have different neutron energy response for thermal and epithermal neutrons. The use of Au and Mn activation measurements in the experimental verification of the epithermal neutron beam is relevant, since the calculated thermal neutron induced dose components form 90–95% of total dose at the depths from 0.5 to 10 cm in a phantom in the FiR 1 beam.
Different MC techniques have been compared to evaluate BNCT dose calculations. The different cross-section forms used in the MCNP and SERA have been shown to cause notable discrepancy in the dose calculation in low (25 meV) and high (1 MeV) energy neutron beams [
], while in the clinical epithermal neutron beam the most crucial neutron-induced boron and nitrogen doses and the photon dose agree within 4% or better [
]. Moreover, discrepancy between the dose calculations may originate rather from the use of different fluence to dose conversion factors or different dose calculation methods chosen in MCNP or in SERA [
The boron concentration level affects directly the boron neutron capture reaction intensity and further the dose to the tumour and the brain. In the Finnish clinical BNCT practice, the whole blood-boron concentration is routinely determined during and after BPA-F infusion from periodic blood samples using inductively coupled plasma-atomic emission spectrometry (ICP-AES) [
]. The time period between the measurements before and after the irradiation can be even 30 min. It has therefore been necessary to find a way to predict the blood-boron level between the measurements and during the irradiation [
Ryynänen P. Kinetic mathematical models for the 111In-labelled bleomycin complex and 10B in boron neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2002.
]. In Finland, a bi-exponential model fit using the Levenberg–Marquardt method has been implemented to provide blood-boron concentration estimates directly for the treatment data flow during the BNCT procedure [
]. The boron concentration in tumour and brain tissues are estimated assuming constant boron concentration ratios for blood-to-tumour (1:3.5) and blood-to-brain (1:1) [
Boron neutron capture therapy of brain tumors: enhanced survival and cure following blood-brain barrier disruption and intracarotid injection of sodium borocaptate and boronophenylalanine.
] measured boron concentration invasively in surgical samples after BPA-F infusion. In patients with suspected or confirmed glioblastoma multiforme, the results of the histological findings indicated mean blood-to-tumour boron concentration ratios from 1.4 to 4 [
]. To date treatment planning codes in BNCT do not factor in the patients' individual characteristics with respect to the varying levels BPA uptake into tumours. Consequently, new non-invasive methods are needed to determine and monitor boron concentrations in the tumour and the surrounding healthy tissues.
Molecular based boron imaging
Many studies have examined the suitability of positron emission tomography (PET) for the determination of boron concentration in tumour and healthy tissues for the needs of treatment planning [
Improved treatment planning for boron neutron capture therapy for glioblastoma multiforme using fluorine-18 labeled boronophenylalanine and positron emission tomography.
]. Such PET studies were conducted using 18F-labelled BPA (18F-BPA). In Finland, 18F-BPA PET has been used to monitor the suitability of patients for BNCT by estimating the accumulation of BPA in the tumour [
]. The 18F-PET has been performed for patients whenever the PET modality and 18F-BPA tracer have been available. The administered activity of 18F-fluoro-l-BPA tracer has typically been 160–240 MBq and tumour-to-normal tissue ratios have been evaluated from static emission scans obtained 20–40 min following injection. In the cases when 18F-PET was performed, the tumour-to-corresponding contralateral normal tissue ratio of the standardized uptake values of at least 2.5 has been required as an inclusion criteria of Finnish head and neck cancer BNCT trial.
Several magnetic resonance (MR) techniques have been proposed for boron carrier detection. These include 11B or 10B magnetic resonance imaging (MRI) and MR spectroscopic methods [
In vivo (19)F MRI and (19)F MRS of (19)F-labelled boronophenylalanine-fructose complex on a C6 rat glioma model to optimize boron neutron capture therapy (BNCT).
]. However, to our knowledge none of these 11B, 10B or 19F MRI/MRS methods for boron detection have been tested in human or implemented in clinical BNCT.
The possibility to apply clinically widely used proton MR spectroscopy (1H MRS) for selective BPA-F detection was proposed by Zuo et al. [
]. They reported that aromatic proton signals of BPA resonating at the chemical shift range of 7.0–7.5 ppm can be detected in vivo at 1.5 T. In theory, this offers possibilities to study the spatial and temporal distribution of BPA and BPA concentration in the brain non-invasively using a typical MR imager without any need for special instrumentation. The aromatic proton signals of BPA enable detection of the boron carrier because, with typical measurement settings (voxel size and measurement duration), normal brain metabolite signals do not overlap with them. MRS imaging (MRSI) of BPA-F with animals at high magnetic field has been studied by Bendel et al. [
Timonen M. Proton magnetic resonance spectroscopy of a boron neutron capture therapy 10B-carrier, L-p-boronophenylalanine-fructose complex. PhD thesis. University of Helsinki, Helsinki; 2010.
Timonen M. Proton magnetic resonance spectroscopy of a boron neutron capture therapy 10B-carrier, L-p-boronophenylalanine-fructose complex. PhD thesis. University of Helsinki, Helsinki; 2010.
]. 1H MRS and/or 3D 1H MRSI measurements were conducted ten patients on the day of BNCT at 1.5 T (nine patients) or 3.0 T (one patient). The patients had glioblastoma multiforme, anaplastic astrocytoma gradus III or head and neck cancer. For nine patients, MRS/MRSI was performed after the treatment, and for one patient MRSI study was started before the end of the BPA-F infusion and ended after the end of the infusion. In three patients, the signals in the aromatic region of tumour spectrum were detected. It appears that in favourable cases, it is possible to detect BPA in vivo after the BPA-F infusion or after the BNCT [
] (Fig. 3 and Fig. 4). However, because the shape and position of the detected signals do not exactly match the BPA spectrum detected in the in vitro conditions, unambiguous assignment of BPA is somewhat difficult. The opportunity to perform MRS immediately after the end of BPA-F infusion for a larger patient group is needed in order to evaluate the actual suitability of 1H MRS for more exact BPA detection and quantification for treatment planning purposes. Using higher magnetic field strengths (3.0 T or even higher) for MRSI the likelihood of detecting low BPA concentrations in vivo would improve. In our work, a 3.0 T MRS study was performed for only one out of ten patients. In this case, the MRS was performed approximately 2 h after BNCT and no signals were detected from tumour spectrum aromatic region. In our previous phantom study [
] 3D acquisition-weighted PRESS (position resolved spectroscopy sequence) MRSI was suggested to be a feasible method for in vivo BPA evaluation in clinical conditions. We applied this method for six patients at 1.5 T and the signals in the aromatic region were detected in the tumour MRSI spectra of two patients. However, use of MRS for exact BPA quantification in vivo would be challenging even if MRS were performed immediately after BPA infusion and modern MRI systems with higher magnetic field were available, but MRSI could be useful for selecting patients for BNCT.
Figure 3The 3D 1H MRSI metabolite maps of a patient with glioblastoma multiforme
Timonen M. Proton magnetic resonance spectroscopy of a boron neutron capture therapy 10B-carrier, L-p-boronophenylalanine-fructose complex. PhD thesis. University of Helsinki, Helsinki; 2010.
. The metabolite maps are based on the Gaussian line shape fitting of two peaks detected in the aromatic region 1.5 h after BNCT. This kind of imaging could give an objective exclusion criterion for BNCT. If no indicative signal is detected from the volume of interest located in a region feasible for good quality MRS, then boron is not accumulating sufficiently in the tumor to reach an appropriate radiation dose level in BNCT.
Figure 4The aromatic region of the 3D 1H MR spectra measured from tumor (A) and contralateral brain tissue (B) 1.5 h after BNCT in patient with glioblastoma multiforme
Timonen M. Proton magnetic resonance spectroscopy of a boron neutron capture therapy 10B-carrier, L-p-boronophenylalanine-fructose complex. PhD thesis. University of Helsinki, Helsinki; 2010.
. In the spectra A and B, baseline correction has been applied. The same data without baseline correction is shown in the spectra C and D for comparison. Two signals can be detected in the tumor spectrum (A). The tumour (E) and contralateral brain tissue (F) spectra measured one month after BNCT and a spectrum measured from 3.0 mm BPA aqueous solution (with 10.1 mm creatine) (G) are also shown.
In 94% of boron neutron capture reactions, a prompt 478 keV gamma photon is created. Measuring this gamma radiation from the patient during the BNCT treatment gives the possibility to monitor and quantify the delivered radiation dose and its distribution during the treatment [
Boron concentrations in brain during boron neutron capture therapy: in vivo measurements from the phase I trial EORTC 11961 using a gamma-ray telescope.
Uptake of two 10B-compounds in liver metastases of colorectal adenocarcinoma for extracorporeal irradiation with boron neutron capture therapy (EORTC Trial 11001).
]. Possibilities to construct a real time 3D imaging device to be used in BNCT treatments have been studied, but is still in the development phase, since the process is demanding for several reasons: the neutrons and other radiation present in the treatment room, the overall low signal-to-noise ratio and the shortness of the treatment and thus imaging time. In particular the prompt 2.2 MeV gamma rays from hydrogen neutron capture reactions with the induced annihilation peak at 511 keV disturb the measurement and the compton edge can give an important contribution to the background on the 478 keV region and must be taken into account.
In 2002, tests with a boron phantom were performed at the Finnish BNCT reactor both with CdZnTe and Si detectors [
]. The CdZnTe detector appeared promising, being able to detect the required boron signal from the background noise. However, the estimated time needed to produce an image was long compared to the treatment time of a patient. Similar results were achieved in Japan, where both HPGe and CdZnTe detectors were tested [
A noninvasive dose estimation system for clinical BNCT based on PG-SPECT–conceptual study and fundamental experiments using HPGe and CdTe semiconductor detectors.
]. Also, optimization of the collimator dimensions was studied, the results indicating that a 1 cm tumour could be detected using an array of detectors [
Boron concentrations in brain during boron neutron capture therapy: in vivo measurements from the phase I trial EORTC 11961 using a gamma-ray telescope.
], but the required space and cooling systems appear to be too extensive to be used in clinical setting. An Argentinean research group has designed and tested a LaBr3(Ce) detector system producing 1 cm3 voxel images with 21 × 21 voxels [
]. With this prototype, spectra from a phantom containing a tumour were measured and an image with 1 cm2 pixels was successfully reconstructed.
During the past ten years, considerable technological improvements have been achieved in both the detector efficiency (electronics, material, signal processing) and shielding of gamma and neutron radiation. At present we are designing and constructing a new prompt gamma camera in Finland. Development of monitoring the dose distribution during the treatment will enable taking BNCT to the next level in terms of predictability and patient safety, as the observation on the exact amount of boron in the tumour as well as in the healthy tissue allows for precise dose calculations.
Diffusion imaging in the detection of BNCT effects in patients
In order to improve BNCT treatments, it is essential to detect the changes in the tumour area and normal tissue as early as possible, and to correlate the findings to the different dose components of the irradiation. MRI and PET studies are used to follow up the response of radiation therapy. Mostly, the follow up of the BNCT patients are based on T2-weighted and gadolinium-enhanced T1-weighted MRI scans. The survival curves are used to determine the “efficiency” of BNCT. However, when a new treatment modality is used for cancer treatment, the acute changes should be especially followed up with modern imaging techniques in order to understand, for example, the effect of different LET-components in the tumour and normal tissues. Conventional imaging techniques seem to be too conservative to detect small changes, which would be essential to understand the mechanisms of complicated irradiations such as BNCT.
Possibilities of using diffusion MRI imaging for tumour grading, definition and response to therapy by estimating changes in Apparent Diffusion Coefficient (ADC) have been widely researched. However, the results appear to overlap and inconclusive for brain tumours [
Tumor extension in high-grade gliomas assessed with diffusion magnetic resonance imaging: values and lesion-to-brain ratios of apparent diffusion coefficient and fractional anisotropy.
Early detection of response to radiation therapy in patients with brain malignancies using conventional and high b-value diffusion-weighted magnetic resonance imaging.
Diffusion-weighted magnetic resonance imaging for predicting and detecting early response to chemoradiation therapy of squamous cell carcinomas of the head and neck.
]. Compared to the typical degree of diffusion weighting (b-value of 800–1000 s/mm2), a higher b-value (e.g. 4000 s/mm²) can yield additional information and thus increase significantly the sensitivity to detect an early response [
Early detection of response to radiation therapy in patients with brain malignancies using conventional and high b-value diffusion-weighted magnetic resonance imaging.
]. Additionally, instead of mono-exponential diffusion analysis, alpha index to model diffusion heterogeneity may be beneficial in assessing the tumour response. However, this has only been reported when assessing the tumour infiltration in rat brain [
]. Functional Diffusion Maps (fDM) is a promising biomarker to predict early response to therapy, whereas the mean ADC for a ROI is insufficient. Pre-clinical animal models suggest that ADC is suitable to predict the efficacy of therapeutic response, while it does not assess the response [
A feasibility study of parametric response map analysis of diffusion-weighted magnetic resonance imaging scans of head and neck cancer patients for providing early detection of therapeutic efficacy.
In Diffusion Tensor Imaging (DTI) directional information about the diffusion is obtained, resulting in anisotropy indices e.g. fractional anisotropy (FA), planar anisotropy (commonly abbreviated as CP) and linear anisotropy (CL) maps. It has been suggested that infiltrative changes can be observed either by visualizing displacement, interruption or widening of the white matter tract in FA-maps or by tractography [
]. FA can be probably used to differentiate tumour recurrence from radiation necrosis in brain tumours and provide a non-invasive alternative to histological diagnosis and craniotomies [
In our preliminary study that aimed to assess the response to BNCT from pre- and post-treatment images from ADC and FA maps, the swelling of the brain caused co-registration problems and the tumour tissue may not have been fully aligned. However, it should be feasible to develop a robust co-registration method once more patients are acquired for the study [
Boron neutron capture therapy (BNCT) followed by intensity modulated chemoradiotherapy as primary treatment of large head and neck cancer with intracranial involvement.
]. The technological aspects of the treatments are a substantial subject of research since standardization of the treatment process and development of patient imaging and monitoring are essential, if the recommendations on strictness and accuracy of the absorbed dose in radiotherapy are to be followed [
Quality assurance in radiotherapy. European society for therapeutic radiology and oncology advisory report to the commission of the european union for the ‘Europe against cancer programme’.
Kosunen A. Metrology and quality of radiation therapy dosimetry of electron, photon and epithermal neutron beams. PhD thesis. University of Helsinki; 1999.
Currently, BNCT beam dosimetry is mainly based on ionization chambers and activation detectors, while other methods such as microdosimeters or gel dosimeters are applied as complementary methods for either absolute or relative dose measurements. Dose measurements still rely partly on results from computer simulations, such as the calculated neutron spectrum at the point of measurement. Thus, the possibilities in research to improve the management and measurement of the epithermal neutron field energy distribution should be continued.
Furthermore, a more detailed knowledge and monitoring of the boron biodistribution would be beneficial for selecting patients for treatments, treatment planning, dose calculations and in estimating the effectiveness and side effects of the treatments. For this purpose, a variety of methods including diffusion imaging, spectroscopy and prompt gamma imaging are under development. The success of BNCT also depends on developing new treatment schemes, such as combining BNCT with photon radiation therapy [
Boron neutron capture therapy (BNCT) followed by intensity modulated chemoradiotherapy as primary treatment of large head and neck cancer with intracranial involvement.
MRI-guided neutron capture therapy by use of a dual gadolinium/boron agent targeted at tumour cells through upregulated low-density lipoprotein transporters.
]. It also seems inevitable that in order incorporate BNCT treatment facilities into hospitals and thus to enable more extensive clinical use of BNCT in the future, development of efficient accelerator based neutron sources is required.
Finally, development in boron neutron capture therapy is an interdisciplinary effort requiring an extensive scientific community with expertise in medical and nuclear physics, reactor engineering, chemistry, pharmacology, clinical radiology, neurology and oncology. To reach the full potential of BNCT, pursuing basic research and technical development should be continued. In Finland, this objective has been promoted over two the last two decades, with work provided by an established research community, professional expertise and dedicated technical infrastructure.
Conflict of interest
None.
Acknowledgements
This work has been supported by the Academy of Finland and the State Subsidy for Helsinki University Central Hospital, Department of Radiology.
References
Joensuu H.
Kankaanranta L.
Seppälä T.
Auterinen I.
Kallio M.
Kulvik M.
et al.
Boron neutron capture therapy of brain tumors: clinical trials at the Finnish facility using boronophenylalanine.
Boron neutron capture therapy of brain tumors: enhanced survival and cure following blood-brain barrier disruption and intracarotid injection of sodium borocaptate and boronophenylalanine.
Boron neutron capture therapy (BNCT) followed by intensity modulated chemoradiotherapy as primary treatment of large head and neck cancer with intracranial involvement.
Seppälä T. FiR 1 epithermal neutron beam model and dose calculation for treatment planning in neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2002.
Järvinen H. Voorbraak W.P. Recommendations for the dosimetry of boron neutron capture therapy (BNCT). The Nuclear Research and consultancy Group NRG,
2003 (Report 21425/03.55339/C)
Boron neutron capture therapy for the treatment of glioblastomas and extracranial tumours: as effective, more effective or less effective than photon irradiation?.
Quality assurance in radiotherapy. European society for therapeutic radiology and oncology advisory report to the commission of the european union for the ‘Europe against cancer programme’.
Uusi-Simola J. Finnish dosimetric practice for epithermal neutron beam dosimetry in boron neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2009.
Study of a solid state microdosemeter based on a monolithic silicon telescope: irradiations with low-energy neutrons and direct comparison with a cylindrical TEPC.
Aschan C. Applicability of thermoluminescent dosimeters in X-Ray organ dose determination and in the dosimetry of systemic and boron neutron capture radiotherapy. PhD thesis. University of Helsinki, Helsinki; 1999.
Kotiluoto P. Adaptive tree multigrids and simplified spherical harmonics approximation in deterministic neutral and charged particle transport. PhD thesis. VTT, Espoo; 2007.
Ryynänen P. Kinetic mathematical models for the 111In-labelled bleomycin complex and 10B in boron neutron capture therapy. PhD thesis. University of Helsinki, Helsinki; 2002.
Improved treatment planning for boron neutron capture therapy for glioblastoma multiforme using fluorine-18 labeled boronophenylalanine and positron emission tomography.
00037-3&id=gr1.jpg){kind=link}
00037-3&id=gr2.jpg){kind=link}
00037-3&id=gr3.jpg){kind=link}
00037-3&id=gr4.jpg){kind=link}