Phys. Med. Bi...
3726                                                                                 K J Riley et al

therapeutic effects ...
A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy            3727

3728                                                                                            K J Riley et al

A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy                           3729

3730                                                                                K J Riley et al

beam monitor counts a...
A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy                             3731

3732                                                                                 K J Riley et al

beam characterizatio...
A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy                         3733

3734                                                                                               K J Riley et al

A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy                         3735

Upcoming SlideShare
Loading in...5

A State of the Art Epithermal Neutron Irradiation Facility for BNCT


Published on

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

A State of the Art Epithermal Neutron Irradiation Facility for BNCT

  1. 1. INSTITUTE OF PHYSICS PUBLISHING PHYSICS IN MEDICINE AND BIOLOGY Phys. Med. Biol. 49 (2004) 3725–3735 PII: S0031-9155(04)81398-8 A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy K J Riley1, P J Binns1 and O K Harling2 1 Nuclear Reactor Laboratory, Massachusetts Institute of Technology, 138 Albany St., Cambridge, MA 02139 USA 2 Department of Nuclear Engineering, Massachusetts Institute of Technology, 138 Albany St., Cambridge, MA 02139 USA E-mail: Received 1 June 2004 Published 2 August 2004 Online at doi:10.1088/0031-9155/49/16/018 Abstract At the Massachusetts Institute of Technology (MIT) the first fission converter- based epithermal neutron beam (FCB) has proven suitable for use in clinical trials of boron neutron capture therapy (BNCT). The modern facility provides a high intensity beam together with low levels of contamination that is ideally suited for use with future, more selective boron delivery agents. Prescriptions for normal tissue tolerance doses consist of 2 or 3 fields lasting less than 10 min each with the currently available beam intensity, that are administered with an automated beam monitoring and control system to help ensure safety of the patient and staff alike. A quality assurance program ensures proper functioning of all instrumentation and safety interlocks as well as constancy of beam output relative to routine calibrations. Beam line shutters and the medical room walls provide sufficient shielding to enable access and use of the facility without affecting other experiments or normal operation of the multipurpose research reactor at MIT. Medical expertise and a large population in the greater Boston area are situated conveniently close to the university, which operates the research reactor 24 h a day for approximately 300 days per year. The operational characteristics of the facility closely match those established for conventional radiotherapy, which together with a near optimum beam performance ensure that the FCB is capable of determining whether the radiobiological promise of NCT can be realized in routine practice. 1. Introduction The first attempt at treating cancers with boron neutron capture therapy (BNCT) was initiated in 1951 (Slatkin 1991) using beams of thermal neutrons. Information about the potential 0031-9155/04/163725+11$30.00 © 2004 IOP Publishing Ltd Printed in the UK 3725
  2. 2. 3726 K J Riley et al therapeutic effects of neutron capture in boron was emerging at that time and seemed to merit clinical investigation despite an incomplete understanding of the complex radiobiological processes that make BNCT so potentially attractive as a modality. The pioneering clinical work of Sweet and colleagues ended when it became clear no therapeutic benefit was being observed and, worse, in some cases excessive damage to superficial normal tissues was being produced (Sweet 1997). BNCT was not tried again in the US until 1994 when separate groups at the Massachusetts Institute of Technology (MIT) and Brookhaven National Laboratory commenced studies for glioblastoma multiforme and subcutaneous melanoma. In the intervening years a rationale for BNCT had become apparent that was based on the preferential accumulation of boron in these tumours using boronated phenylalanine (BPA) (Mallesch et al 1994, Coderre et al 1998). Important technological advances had also occurred to allow rapid assay of specimens containing boron by prompt gamma neutron activation analysis (PGNAA) and more importantly the design and construction of epithermal (0.5 eV–10 keV) neutron beams that would avoid excessive skin dose and provide a therapeutic effect on deep-seated tumours. Interest in BNCT was renewed and facilities around the world soon launched Phase I clinical studies to determine the maximum tolerable dose using these epithermal beams in conjunction with new boron compounds (Chanana et al 1999, Busse et al 2003). The use of research reactors was the only pragmatic solution to provide a reliable and intense source of neutrons suitable for NCT although their design was not necessarily compatible for producing an optimum beam of sufficient intensity and purity. Many BNCT facilities have low beam intensities that result in protracted irradiations, higher levels of contamination causing unnecessary damage to healthy tissue that degrade treatment plans or poor physical designs that constrain field placement. These shortcomings, although a serious impediment, have so far not greatly hindered BNCT because the trials to date have simply focused on normal tissue tolerance and not efficacy or tumour control. BNCT still requires optimization of the complex biological and physical factors that distinguish it from conventional radiotherapy and that will only be possible after Phase II clinical investigations. The fission converter concept is employed for the first time to provide a second-generation epithermal neutron beam that is nearly optimum for clinical trials and successfully overcomes the difficulties of extracting a suitable beam from the reactor core. The facility (FCB) complies with all of the envisioned requirements for rigorous clinical evaluation of BNCT (Harling et al 2002a, Riley et al 2003) and is housed in the experimental hall of the MIT Research Reactor (MITR-II). The reactor is located in metropolitan Boston with several teaching hospitals nearby in the densely populated New England region of the US. Patients are admitted to the hospital and travel to the reactor by ambulance for part of the day to receive irradiations and then return to the clinic for 2–3 days of in-patient care and follow-up. The multipurpose MITR-II operates at a maximum of 5 MW, 24 h a day for approximately 300 days per year and the FCB runs independently of other experiments and does not interfere with other reactor applications. The FCB provides a high purity, high intensity beam that enables irradiations to average whole brain tolerance in approximately 25 min using boron compounds currently approved by the Food and Drug Administration (FDA). The facility is augmented by the PGNAA (Riley and Harling 1998) on a separate beam line in the reactor hall to enable rapid assay of the boron content in biological specimens from the patient that is necessary to adjust beam delivery to achieve the specified treatment plan (Kiger et al 2003) with a precision better than 2%. Together these facilities are used in the current clinical trials for glioblastoma multiforme and intracranial as well as metastatic melanoma and are the only facilities licenced for BNCT in the US.
  3. 3. A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy 3727 Figure 1. Plan view of the MIT epithermal neutron fission converter facility. 2. Facility description The fission converter is designed and licenced to operate at 250 kW and is presently configured to produce 83 kW by using only 10 partially burned fuel elements cooled with D2O. A shielded horizontal beam line 2.5 m long directs neutrons from the converter to the treatment room as shown schematically in figure 1. The beam line consists of a series of Al (81 cm), Teflon R (13 cm) and Cd (0.5 mm) neutron filter/moderators, a lead photon shield (8 cm), and a large conical collimator 1.1 m long with lead walls 15 cm thick. The 0.42 m long patient collimator is made from a mixture of lead and boron or lithium (95% enriched in 6Li) loaded epoxy that extends the beam line into the shielded medical room. The neutron beam is controlled by three in-line shutters acting independently that are installed along the length of the beam line. The first of these, starting near the reactor core is the converter control shutter (CCS) that is a 0.5 mm layer of Cd followed by a 6.4 mm sheet of aluminium alloyed with boron of natural isotopic abundance. This shutter modulates the fission rate in the converter (and beam intensity) between 1 and 100% by shielding the converter fuel against thermal neutrons incident from the MITR-II reflector region. Downstream from the fuel is a 68 cm long tank that when filled with light water provides effective neutron attenuation. Following this is the mechanical shutter that turns the beam on and off within 3 s during therapy and comprises a large sliding slab to fill a section of the collimator with a 20 cm thickness of borated (100 mg cm−3 10B) high density (ρ = 4.0 g cm−3) concrete and 20 cm of lead. The medical room is built with 1.1 m thick walls of high density concrete with a roof of 15 cm thick steel beneath 55 cm of high density concrete. The wall of the medical room nearest the FCB control console includes a window containing layers of quartz and lead glass as well as mineral oil. Inner surfaces of the walls and ceiling are lined with 2.5 cm of borated polyethylene to absorb thermal neutrons and reduce activation of the steel reinforced concrete walls. The beam centreline in the medical room is 0.42 m above the floor and the patient collimator can be readily configured to provide aperture diameters of 80, 100, 120 and 160 mm
  4. 4. 3728 K J Riley et al Figure 2. Setup for a lateral brain irradiation using the patient collimator (1) with the 120 mm diameter aperture (2). Clinical staff monitor the patients and their vital signs (3) during an irradiation through the shielded viewing window (4) and closed circuit cameras (5). that conveniently extend up to 0.42 m beyond the wall of the medical room. The collimator diameter tapers from 0.67 m at its base to 0.3 m near the patient that combines with ample (14 m2) floor space in the medical room to allow patients to be comfortably positioned for cranial irradiations in a full 180◦ arc around the beam centreline while lying supine on the treatment couch. A photograph of a patient setup for a lateral brain irradiation is shown in figure 2. A laser projection illuminates the central axis of the beam to help with patient positioning as well as optics that penetrate the wall of the collimator to provide a beam’s eye view. Prior to commencing an irradiation, the laser and optics are withdrawn and replaced with a plug that has a composition identical to the collimator walls. Four fission counters positioned in the periphery of the beam near the base of the patient collimator serve as integral monitors of the neutron fluence as it is delivered to the patient. Signals are fed to NIM electronics and irradiations are administered with a programmable logic controlled (PLC) system that automatically terminates the irradiation when the integrated counts on any of the four beam monitors first reach the prescribed targets. Data from instrumentation in the FCB cooling system and beam line shutters are also fed to the PLCs, which are programmed with automated interlocks to help ensure the safety of the patient and operational staff alike. The facility is operated from the control console shown in figure 3 that includes a dedicated computer for displaying progress of an irradiation and archiving data from the PLCs. During an irradiation the patient and their vital signs are monitored through the shielded viewing window and closed-circuit cameras which contain an integrated
  5. 5. A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy 3729 Figure 3. The FCB control console is equipped with two redundant PLCs (1) that automatically control an irradiation. Signals from the four beam monitors are processed in NIM electronics (2), fed to the PLCs then stored and displayed on a computer (3). The console is equipped with closed-circuit television monitors (4) and is situated near the shielded window (5) so that patients can be easily viewed throughout the course of an irradiation. Key switch on Enter beam Satisfy safety monitor targets interlocks Medical room door closed Power for shutter systems BEV laser removed Open shutters Reset NO Interlocks OK? YES Log data Beam Monitors less than targets? NO (every 10 ms) monitors YES Operator intervene SCRAM YES NO Over targets? Close shutters (> 102%) Figure 4. Logic diagram of the FCB control system. audio system for two-way communication between the medical room and control console. The control system, medical room and support for the patient at the FCB are designed to function as similarly as possible to conventional radiotherapy facilities in terms of safety and convenience. 3. Beam monitoring and control system 3.1. Overview The control logic during a patient irradiation is depicted in figure 4. Prior to commencing an irradiation, a series of safety interlocks must be satisfied before the shutters can be opened to turn the beam on. The prescribed monitor units are adjusted based upon PGNAA measurements of the boron concentration in blood samples taken prior to irradiation and are entered using a numeric keypad on the console. To prevent input errors, the system only accepts the entered
  6. 6. 3730 K J Riley et al beam monitor counts after matching values for each channel are keyed in twice. The operator commences therapy with a single pushbutton and the PLCs issue commands to open each shutter in sequence and initiate data acquisition. It takes two minutes for all shutters to open, and monitor counts accumulate continuously on the updated display. The PLCs repeatedly interrogate all safety interlocks, check that the accumulated monitor counts are below the pre- set targets and store the data from that interval in the computer. Like conventional radiotherapy machines no other actions are required from the operator unless they need to intervene, for which there is a manual override that terminates an irradiation by closing shutters or scramming the reactor. When the accumulated counts on any one of the four beam monitors first reaches the set target, the PLCs signal all shutters to close. To defend against overexposures that might be caused by some mechanical or electrical failure during shutter closure, programmed safety interlocks automatically scram the reactor if any channel exceeds 102% of the prescribed target value. Controls for opening shutters are deactivated when the shield door to the medical room is open to help prevent inadvertent beam exposure of staff inside the room. Additional shutter closure controls mounted inside the room are always active. If radiation levels inside the room exceed 0.5 mSv h−1 an audible alarm warns personnel upon opening the door. The entrance to the medical room is equipped with motion sensors that stop sideways movement of the pneumatically operated 11 ton shielding door if anyone is in the vicinity and pressure sensitive strips run along its leading edge to stop the door upon any contact. Loss of building power would automatically scram the MITR-II, but if electrical power fails only to the medical area, uninterruptible power supplies keep the PLCs, computer and other vital instrumentation running for at least 20 min to enable an irradiation field to be completed as planned. The mechanical shutter can be rapidly closed using a hand crank located on the outside of the room while the water shutter and CCS close under the force of gravity. The shielding door can also be opened by hand in an emergency to quickly gain access to the medical room. 3.2. Performance Seven clinical irradiations have so far been completed using the FCB without experiencing any unexpected difficulties except for one reactor scram that was unrelated to operating the medical facility and resulted in only a 1 h delay in giving the next radiation field. Average whole brain doses of either 7.0 or 7.7 (RBE) Gy were delivered, with total irradiation times that ranged from 25 to 79 min. Clinical irradiations commenced with the converter operating at approximately 30 kW, which resulted in proportionately longer irradiations. Treatment plans consisted of two or three fields that each lasted between 2 and 15 min and were delivered on two consecutive days. Each of the 36 separate fields administered using the FCB were delivered to within 1.5% of the prescribed targets, averaged over the four beam monitors (Kiger et al 2004), with 33 (92%) and 21 (58%) fields delivered to within 1 and 0.5%, respectively. 4. Health physics surveys Photon and neutron dose equivalent rates were determined using Geiger–M¨ ller and tissue u equivalent proportional counter based survey instruments, respectively. With the reactor operating at 4.5 MW, measurements were performed inside the medical room with the shutters closed and outside the room with the shutters open (FCB operating at 76 kW) and scaled to maximum operating levels (reactor power of 5 MW, converter power of 83 kW). The dose equivalent rates outside the medical room contained no significant neutron contribution with
  7. 7. A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy 3731 Table 1. Summary of the quality assurance programme for radiotherapy at the MIT FCB. Surveillance Interval Criteria Functional checks Safety interlocks Monthly Functional Alarm setpoints Each use Functional, set according to TS Controls Each patient Functional Equipment Each patient Functional Output/constancy checks Area radiation monitors Each patient Alarm operational Beam monitors (voltage and Semi-annual Set point stable within 10% discriminator plateaux) Boron analysis system Each patient Constant within 10% Ionization chambers, electrometer Each use Constant within 3% Neutron beam Each patient Constant within 10% Calibrations Area radiation monitors Annual 10% accuracy Boron analysis system Semi-annual 5% accuracy Detector efficiency (gold foils) Semi-annual 3% accuracy FCB power monitors Annual – Ionization chambers, electrometer Bi-annual National standard, 1% precision Neutron beam depth dose profile Semi-annual (each – beam configuration) a maximum of 12 µSv h−1 measured behind the rear wall opposite the beam. Since the observed values are only marginally higher than the nominal background of approximately 8 µSv h−1 in the reactor hall without the converter operating, no additional access control to the experimental hall is required when the FCB is in use. Inside the medical room a dose equivalent rate of 200–300 µSv h−1 is apparent at the patient position immediately following a 10 min irradiation that is due entirely to photons emanating from the beam line. Approximately half of this activity arises from relatively short- lived activation products (28Al, 56Mn and short-lived isomers of 122Sb and 124Sb) produced in the patient collimator and after 12 h the dose equivalent rate is decreased to 100 µSv h−1. This is higher than the 26 µSv h−1 associated with beam transmission through the closed shutters that was first measured before any use of the facility and is attributed to a small build up of residual activity in the patient collimator during its lifetime. General area dose rates of approximately 20 µSv h−1 are observed away from the patient collimator in the medical room with the reactor at full power and staff can therefore freely enter the room without the need to lower reactor power. 5. Quality assurance programme A set of procedures was developed and is routinely performed to ensure proper functioning of all systems and to maintain accurate calibration of beam monitor output in terms of the absorbed dose delivered under reference clinical conditions. These procedures parallel those used in conventional radiotherapy and were designed to fulfil technical specifications (TS) for the FCB as well as the quality assurance programme for the conduct of human therapy approved by the US Nuclear Regulatory Commission. The procedures summarized in table 1 are separated into three different categories; functional checks, output or constancy checks, and
  8. 8. 3732 K J Riley et al beam characterizations or instrument calibrations. Compliance is documented by archiving completed forms and a master schedule that tracks each procedure and indicates the next necessary completion date. Integral functional checks of all safety-related interlocks are performed on a monthly basis. Proper functioning of equipment (e.g. temperature and flow monitors, nuclear instrumentation, etc) is verified prior to each use of the facility through a system start-up checklist. Prior to commencing clinical irradiations, an additional checklist is completed to ensure that patient- related equipment such as cameras, intercoms and door controls are operational. Output or constancy checks of the neutron beam, area radiation monitors, boron analysis system and other detectors are generally completed as needed prior to either their use or a clinical irradiation. The discriminator and voltage plateau checks for each beam monitor are performed every six months. Each beam configuration (e.g. beam aperture) must be characterized within six months prior to its clinical use. A beam characterization consists of measurements along the central axis of an ellipsoidal, water filled phantom to determine the absorbed dose under reference clinical conditions. The results from these measurements are correlated with output from the beam monitors and the treatment planning system so that irradiations can be administered using a set of monitor units that are separately prescribed for each of the four beam monitors. Instrumentation used during the characterization must also be calibrated. Ionization chambers are calibrated against a national standard in terms of air kerma for 60Co at an accredited dosimetry calibration laboratory every two years and constancy checks are performed on an in-house photon source prior to each use. The boron analysis systems as well as the area radiation monitors inside and outside the medical therapy room are also calibrated at regular intervals. 6. Beam data The technique of mixed-field dosimetry described previously for epithermal neutron beams (Rogus et al 1994, Riley et al 2003) is applied to measure the neutron and photon absorbed dose rates inside a 0.6 × 0.6 × 0.6 m3 water phantom with 10 mm thick PMMA walls. The dimensions of the phantom are much greater than the 160 mm diameter of the largest aperture to minimize the effects of leakage radiation for measurements near the edge of the field. The phantom was positioned with the front face against the end of the patient collimator without an air gap. Depth dose profiles were measured along the central axis and at a horizontal displacement of 80 mm from the central axis of the beam. Horizontal cross-field profiles (displacement ranging from −100 to 100 mm) were obtained at a depth of 25 mm near the total dose maximum. These data complement the characterization measurements routinely performed as part of the quality assurance program and serve to benchmark the source term used for treatment planning calculations. These data are also being used for the international programme combining clinical results from BNCT centres worldwide (Harling et al 2002b). The measured photon absorbed dose rate has an estimated uncertainty (1σ ) of 4.4%, while that attributed to thermal neutrons ranges from 4.6% near the surface of the phantom to 6.5% at depth due to differences in counting statistics of the activation foils. Fast neutrons account for only approximately 5% of the total measured response of the A-150 walled ionization chamber and consequently the large correction necessary in the twin chamber method limits the accuracy of determination. Fast neutron absorbed dose rates near the surface have an estimated uncertainty of 61%, while those at depths of 50 mm and greater are 125%. The depth dose profiles measured on the central axis of the phantom and at a lateral displacement of 80 mm near the edge of the field are shown in figures 5(a) and (b), respectively,
  9. 9. A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy 3733 (a) (b) (c) ◦ Figure 5. Absorbed dose rates for photons ( ), thermal neutrons ( ) and fast neutrons (♦) measured in a 0.6 × 0.6 × 0.6 m3 water phantom for the 160 mm diameter field and scaled to a converter operating power of 83 kW with depth-dose profiles (a) on the central axis of the beam (b) displaced 80 mm from the central axis and (c) cross-plane profiles measured at a depth of 25 mm with the aperture indicated by dashed lines.
  10. 10. 3734 K J Riley et al scaled to a converter operating power of 83 kW. The shallowest measurements on the depth dose curve are near the expected dose maximum for the photon and thermal neutron dose components, which exhibit a steady decrease with depth in phantom. The measured fast neutron absorbed dose rates are comparatively small, particularly at depth in phantom and for the off-axis profile. Figure 5(c) shows cross-plane profiles of the different dose components measured at 25 mm depth in the phantom. All of the dose components decrease steadily with increasing displacement and appear symmetric about the central axis of the beam. The 80–20% penumbrae for the different dose components are large (>40 mm) compared to other modalities where the spatial profile of the beam must be carefully tailored to match the boundaries of the planning target volume. In BNCT, a uniform distribution of thermal neutrons within the target volume is desirable because without significant levels of beam contamination, as exhibited by the FCB, the dose will be principally confined to tumour regions where the boron concentration is expected to be highest. 7. Summary The fission converter concept has proven suitable for obtaining a high purity beam of epithermal neutrons for BNCT with intensities that result in irradiation times as short as a few minutes. The relatively low power (83 kW) generated in the converter illustrates the efficiency of the fission process for producing epithermal beams and the feasibility of small reactor-based sources for dedicated use in a hospital. Though the MITR-II is not dedicated solely for BNCT research, the FCB and PGNAA are independent of other experiments and do not affect regular reactor operation. Several nearby New England hospitals can conveniently conduct outpatient irradiations at MIT, drawing from a large local population. The beam line is presently optimized for brain tumour studies although it can be easily reconfigured to treat other disease sites. The operational characteristics of the facility closely match those established for conventional radiotherapy, which together with a near optimum beam performance ensure that the FCB is capable of determining whether the radiobiological promise of this cellular tumour targeting therapy can be realized in routine practice. Acknowledgment This work has been supported by the US Department of Energy under contract number DEFG02-96ER62193. References Busse P M et al 2003 A critical examination of the results from the Harvard-MIT NCT program phase I clinical trial of neutron capture therapy for intracranial disease J. Neuro-Oncol. 62 111–21 Chanana A D et al 1999 Boron neutron capture therapy for glioblastoma multiforme: interim results from the Phase I/II dose-escalation studies Neurosurgery 44 1182–93 Coderre J A, Chanana A D, Joel D D, Elowitz E H, Micca P L, Nawrocky M M, Chadha M, Gebbers J O, Shady M and Slatkin D N 1998 Biodistribution of boronophenylalanine in patients with glioblastoma multiforme: boron concentration correlates with tumor cellularity Radiat. Res. 149 163–70 Harling O K et al 2002a The fission converter-based epithermal neutron irradiation facility at the Massachusetts Institute of Technology Reactor Nucl. Sci. Eng. 140 223–40 Harling O K et al 2002b International dosimetry exchange: a status report Research and Development in Neutron Capture Therapy ed W Sauerwein, R Moss and A Wittig (Bologna: Monduzzi) pp 333–9 Harling O K and Riley K J 2003 Fission reactor neutron sources for neutron capture therapy—a critical review J. Neuro-Oncol. 62 7–17
  11. 11. A state-of-the-art epithermal neutron irradiation facility for neutron capture therapy 3735 Kiger W S III et al 2004 Preliminary treatment planning and dosimetry for a clinical trial of neutron capture therapy using a fission converter epithermal neutron beam J. Appl. Radiat. Isotopes at press Kiger W S III, Palmer M R, Riley K J, Zamenhof R G and Busse P M 2003 Pharamacokinetic modeling for boronophenylalanine-fructose mediated neutron capture therapy: 10B concentration predictions and dosimetric consequences J. Neuro-Oncol. 62 171–86 Mallesch J L, Moore D E, Allen B J, McCarthy W H, Jones R and Stening W A 1994 The pharmacokinetics of p-boronophenylalanine fructose in human patients with glioma and metastatic melanoma Int. J. Radiat. Oncol. Biol. Phys. 28 1183–8 Riley K J and Harling O K 1998 An improved prompt gamma neutron activation analysis facility using a focused diffracted neutron beam Nucl. Instrum. Methods B 143 414–21 Riley K J, Binns P J and Harling O K 2003 Performance characteristics of the MIT fission converter based epithermal neutron beam Phys. Med. Biol. 48 943–58 Rogus R D, Harling O K and Yanch J C 1994 Mixed field dosimetry of neutron beams for boron neutron capture therapy at the MITR-II research reactor Med. Phys. 21 1611–25 Slatkin D N 1991 A history of boron neutron capture therapy of brain tumors Brain 114 1609–29 Sweet W H 1997 Early history of development of boron neutron capture therapy J. Neuro-Oncol. 33 19–26