- Email: [email protected]
Enhancement of the epithermal neutron beam used for boron neutron capture therapy
Int. J. Radiation
Oncology
Pergamon
Biol. Phys., Vol. 28, No. 5, pp. 1149-I 156, 1994 Copyright 0 1994 Elsevier Science Ltd Printed in the USA. All rights reserved 0360-3016/94 $6.00 + .OO
0360-3016(94)E0076-V
??Special Feature- Physics Original Contribution
ENHANCEMENT OF THE EPITHERMAL NEUTRON BEAM USED FOR BORON NEUTRON CAPTURE THERAPY HUNGYUAN DAVID
B. Lru,
C.
RORER,
*Medical
PH.D.,*
Department
M. BRUGGER,
ROBERT
PH.D.,?
JIH-PERNG
and tReactor
PH.D.,*
DENNIS D. GREENBERG, B.S.,*
Hu, PH.D.-~ AND HENRY M. HAUPTMAN,
Division,
Brookhaven
National
Laboratory,
Upton,
B.S.-j-
NY 11973
Purpose: This report describes a study to enhance the epithermal neutron beam at the Brookhaven Medical Research Reactor by increasing the epithermal neutron flux and/or reducing contamination by fast neutrons. Methods and Materials: The beam was reevaluated using Monte Carlo calculations and flux and dose measurements in air and in an ellipsoidal head phantom at the patient irradiation port. Changes in its geometry and materials were considered, including rearranging the fuel elements in the reactor core and redesigning the moderator and the patient irradiation port. Results: Calculations of the new fluxes and doses at the patient irradiation port showed that the epithermal neutron flux can be increased by lOO%, while the fast neutron dose per epithermal neutron can be reduced by 38%. In 1992, some of the proposed changes were made. In June 1992, measurements were made after one additional fuel element was added to replace a graphite spacer block on the epithermal beam side of the reactor core. The results show that the epithermal neutron flux increased by 18%, as predicted by the Monte Carlo calculations. In October 1992, the fuel elements in the reactor core were rearranged by placing four new fuel elements in the first row facing the epithermal port; the intensity of the epithermal neutron beam increased by 50% and the fast neutron and gamma doses per epithermal neutron may have decreased slightly. Conclusion: The epithermal neutron beam at the Brookhaven Medical Research Reactor has gained a 50% increase in the epithermal neutron flux and the fast neutron and gamma doses per epithermal neutron are reduced slightly after the rearrangement of the fuel elements in the core. Boron neutron capture therapy, Brookhaven dosimetry.
Medical Research
Epithermal
neutron beam, Neutron
lignant tissue while the dose imposed on the healthy tissue is below the threshold for irreversible damage. The initial clinical trials of BNCT in the 1950s at the Brookhaven National Laboratory (BNL) were made with a beam of thermal neutrons from the Brookhaven Graphite Research Reactor (BGRR) (8). The Brookhaven Medical Research Reactor (BMRR) (9) was built and became operational in 1959 to provide a higher flux beam of thermal neutrons for BNCT research. After the initial trials, it was realized that to reach deep-seated brain tumors, a beam of epithermal neutrons was preferable to a beam of thermal neutrons. The depth of penetration of a beam of thermal neutrons is limited, so excessive damage to the skin and crania1 structures occurs. In contrast, epithermal
INTRODUCTION
Optimal cancer therapy would eliminate tumor cells without seriously damaging the normal tissue. Today’s standard treatments have succeeded in curing some kinds of cancers, but for others, such as malignant brain tumors, we have achieved only marginal lengthening of the patient’s life. New treatment modalities are needed to selectively irradiate the tumor cells while sparing the normal tissue. Boron neutron capture therapy (BNCT) (2) is based on a two-part modality to treat these resistant tumors. It brings together two components, namely l”B nuclei and thermal neutrons, that, when used separately, have only minor effects on the normal cells. The combination of these two components at the tumor site releases intense radiation of (Y and Li particles that can destroy the ma-
Data presented, in part, at the Fifth International
Reactor (BMRR),
neutrons, keV,
Symposium
have
those
having
a greater
energies
depth
between
of penetration
0.4 eV and so that
10
a deep-
We thank C. R. Gordon and L. S. Warkentien for their handling of the cell survival measurements and A. L. Ruggiero for clerical assistance. We also thank A. D. Woodhead for final editing of this manuscript. This research was supported by the US Department of Energy under contract DE-AC02-76CHOOO 16 with Brookhaven National Laboratory. Accepted for publication 2 September 1993.
on Neutron Capture Therapy, Columbus, OH, 13- 17 September 1992. Reprint requests to: Hungyuan B. Liu, Ph.D., Medical Dept., Building 490, Brookhaven National Laboratory, Upton, NY 11973. Acknowledgement-We thank J. Gajewski for assistance preparing the geometrical design for the Monte Carlo calculations. 1149
1150
1. J. Radiation Oncology 0 Biology 0 Physics
seated tumor can be treated while sparing surface tissue. In 1988, a moderator was inserted into one of the beam shutters at the BMRR to produce a beam of epithermal neutrons. This beam, which has a relatively low contamination of fast neutrons. has been shown to be satisfactory for BNCT. A patient can be fully treated in about an hour. As a practical matter, the ability of patients to tolerate the psychological stress associated with lying motionless in a neutron beam for an hour is a concern. Therefore, to shorten the treatment time, it is worthwhile to increase the epithermal neutron flux. Also, while the present BMRR epithermal neutron beam is satisfactory for BNCT, it would be more effective if the fast neutron dose could be reduced to spare the skin more. The purpose of this study was to show how the epithermal neutron beam at the BMRR can be improved by increasing the flux of epithermal neutrons and reducing the fraction of fast neutrons. We used Monte Carlo calculations to predict the changes in fluxes produced by changes in the material or geometry in the core of the BMRR and in the moderator of the beam shutter and the patient irradiation port. Several of the proposed changes were made in 1992 and new measurements are compared to earlier ones.
BACKGROUND
OF
BNCT
AT BNL
Eurly triuls qf BNCT Initial clinical trials of BNCT began in 195 1 using a thermal neutron beam at the BGRR. By removing part of the concrete shielding over the top of the BGRR, an irradiation port was built to deliver thermal neutrons to irradiate brain tumors in patients. Immediately after the intravenous injection of a borax solution, the patient’s head was fixed in the proper position over the irradiation port. The reactor was then brought to full power to irradiate the patient for about half an hour after which the reactor was shut down. The patients were irradiated by thermal neutrons with the inevitable contamination by fast neutrons and gamma rays. During the 195Os, 28 patients were treated at the BGRR. The BMRR became operational in 1959 to deliver a more intense beam of thermal neutrons for BNCT, with an improved beam-extraction facility that facilitated future beam development. Between 1959 and 1961, 17 patients were treated at the BMRR. Analysis of the results of these early clinical trials showed they had little significant therapeutic effect (12). One patient in this series of BNCT trials did not deteriorate neurologically after treatment; however, he died 5 months after treatment from widespread extracranial metastases. At necropsy, there was no evidence of viable tumor tissue at the primary site. Concurrently, efforts toward developing BNCT were also made at the Massachusetts Institute of Technology Reactor (MITR). The outcome of the MIT clinical trials of BNCT was also unsatisfactory ( 1). Failure to demonstrate substantial extensions of lifespan resulted in suspensions of both the BNL and the MIT clinical trials of BNCT.
Volume
28, Number
5, 1994
No patients have been treated by BNCT in the United States since 196 1. The experimenters concluded that the ‘“B compound did not concentrate and persist in the tumor cells, and that the thermal neutrons did not penetrate to the adequate depths. Therefore, there was excessive damage to the normal tissue and inadequate destruction of the deep-seated brain tumor. Better “B compounds that target tumors were needed together with beams of epithermal neutrons that would penetrate to deep-seated tumors. BMRR The BMRR is an adjunct facility to the BNL’s Medical Research Center, and is located in an 18.3 m-diameter gas-tight confinement building which is connected to the laboratory through an air lock. Figure 1 shows the cutaway configuration of the BMRR. There are two treatment neutron beam ports, each within a shielded room, on opposite faces of the reactor. At present, one is a thermal neutron beam and the other is an epithermal neutron beam. The emerging neutron beams at the patient irradiation ports are interrupted by beam shutters with assemblies that can be raised and lowered hydraulically inside a vertical cavity to control the irradiation. When the shutter is down, a high-density concrete section of each shutter blocks the beam between the reactor core and the patient location. When the shutter is raised, the moderator section of the shutter is between the reactor core and the patient location. When the moderator is D20, a thermal neutron beam is produced and when the moderator is Al and A1203, an epithermal neutron beam is produced.
Fig. 1. Cut-away view of the BMRR, showing (A) the reactor core, (B) the beam shutter, and (C) the moderator in the beam shutter.
Boron neutron capture therapy 0 H. B.
II
within the Al vessel is filled with graphite or Al blocks. The control system consists of three safety rods and one regulating rod. An air-cooled reflector structure of graphite blocks surrounds the reactor core. Two stationary Bi block walls in the beam’s direction within the reflector were added to shield patients from gamma rays from the reactor core. Figure 2a depicts a horizontal section of the central plane of the BMRR extending from the reactor core out to the epithermal patient irradiation port, showing the reactor core, graphite reflector, inner Bi shield, moderator tanks in the A, B, and C regions, beam shutter, outer Bi shield, and Li2C03-in-polyethylene (Li-poly) shield at the patient irradiation port. The irradiation points at the irradiation port are labeled X and Y. In 1959, the reactor started up with 17 new fuel elements; with time. the fuel was burned up, but the partially burned elements were not removed and new elements were added until May 1992 when there were 30 partially spent fuel elements in the reactor core. In June 1992, one extra fuel element was added to the reactor core in place of a graphite spacer block. The reactor core was reconfigured in October 1992 and there are now 3 1 fuel elements in the reactor core with four new fuel elements facing the epithermal irradiation port. The moderator region C consists of two empty spaces with areas of 127 X 119 cm2 and 4 and 8 cm thick, respectively. These empty spaces can be filled with liquid or pellets. The shutters are 1.9 cm-thick steel shells, filled with high density concrete at the top and the moderator tanks A and B near the bottom. The Al and A1203, which were identified by Brugger and Less (5) as good moderators to produce an epithermal neutron beam, were selected by Fairchild and Wheeler (7) as the primary moderator and used to fill in the A and B regions of the existing beam shutter to produce the present epithermal neutron beam. Figure 2b shows the A, B, and C regions of the present neutron moderator. The moderator is arranged in an alternate sequence of Al and AlzOj, followed by the Bi shield both in the beam shutter and at the patient irradiation port. A 3.8 cm-thick Li-poly shield was added in 199 1 around the Bi port to reduce the neutron flux coming from outside the port.
High Density Li-Poly Shield
Gmcrcte
J -Bi
D
1Pbl +
II
a
Li-Poly
I I I Shic’d
Beam’Sbutter
Li-Poly Shield
LiPoly Shield
b Fig. 2. (a) Horizontal section of the present BMRR epithermal neutron beam, showing the moderator tanks in the A, B, and C regions. (b) Horizontal section of the present BMRR epithermal neutron beam, showing the mixing assembly of Al and A&O3 in moderator tanks A and B.
The reactor core is cooled by an upward flow of water and operates at a maximum power of 3 MW. Most components of the reactor core, including the reactor vessel and cooling water system, are made of Al. Space not occupied by fuel elements, control rods, or instrumentation
Table
BMRR at BMRR at BMRR at BMRR at MITR* PETTEN+
X Y X Y
(May (May (Ott (Ott
1992) 1992) 1992) 1992)
* Personal communication, MA, USA. + Personal communication,
1. Epithermal
neutron
Power (MW
Epithermal neutron flux lo9 (n/cm* set)
3 3 3 3 5 45
1.8 1.2 2.7 1.8 0.20 0.33
Dr. 0. Harling,
Nuclear Engineering
Dr. R. Moss, Commission
1151
LIU et a/
beams for BNCT Fast neutron dose per epi-neutron lo-” cGycm*/n
Gamma dose per epi-neutron 10-i ’cGycm*/n
4.8 4.8 4.3 4.3 13 10.4 Department,
ofthe European
Massachusetts
Communities,
1.1 1.4 1.0 1.3 14 8.4 Institute
of Technology,
Cambridge,
Joint Research Centre, Petten, The Netherlands.
1152
I. J. Radiation Oncology 0 Biology 0 Physics
Since the beam moderator was assembled in 1988, dosimetric measurements have been carried out in air and in phantoms at the epithermal irradiation port (3, 10, 11) to define the incident flux and the imparted dose as a function of energy in the epithermal neutron beam. Compared to the other two operating epithermal neutron beams, one in Cambridge, MA and one in Petten, The Netherlands, (6,13) the BMRR epithermal neutron beam delivers the highest epithermal neutron flux and the lowest fast neutron dose per epithermal neutron. Table 1 compares the parameters of these epithermal neutron beams. Liu (10) calculated treatment plans for treating brain tumors by irradiation from the BMRR epithermal neutron beam. In one such calculation, the patient is placed at Y (Figures 2a and 2b), and is irradiated with the BMRR epithermal neutron beam at 3 MW power, with a 10 ppm “B uniformly distributed through the head as a baseline concentration, an irradiation of 54 min is required to reach a dose tolerance limit of 10 RBE-Gy delivered to the center of the brain ( 12). The RBE-Gy stands for the relative biological effectiveness (RBE) of a radiation component multiplied by the physical dose of the component (Gy). Compared to traditional external beam radiation therapy which usually is completed in a few minutes per fractionation, a reduction of the 54-min irradiation time at the BMRR becomes important to ameliorate the patient’s anxiety. In addition, a reduction of the fast neutrons in the beam is also important to reduce the dose to the patient’s skin, since the fast neutron RBE dose contributes over 50% of the dose to the skin.
METHODS
AND
MATERIALS
To optimize a physical facility, a reliable calculation technique is needed to predict the outcome of a change of the design, with dependable experimental methods to validate this change. In this study of upgrading the BMRR epithermal neutron beam, a Monte Carlo code was used to predict the improvements, and then, where feasible, dosimetric measurements in air and in an ellipsoidal head phantom were made and repeated to substantiate the Monte Carlo predictions.
Volume 28, Number 5, 1994
version, or escape. Then, another particle is sampled to start the interaction with materials all over again and so on, for thousands to millions of particles. The Monte Carlo for Neutron and Photon (MCNP) transport code, which was developed at the Los Alamos National Laboratory (4) is a general-purpose Monte Carlo code, which can be used to model the transport of neutrons, photons, coupled neutrons/photons, or coupled photons/electrons (MCNP version 4.2). The code treats an arbitrary 3-D configuration of materials in geometric cells bounded by planes, spheres, cones, or ellipsoids. MCNP uses either continuous or discrete nuclear crosssection libraries to describe the probability of reactions. Nuclear data is tabulated for neutron interactions, photon interactions, thermal particle scattering, and neutron dosimetry and activation. The code is well designed and experimentally validated, so it is a good computational tool for applications in the neutron beam designs of BNCT. A detailed MCNP input program was designed to simulate different areas of interest for the BMRR core and the epithermal neutron beam facility. MCNP (version 4.0) calculations were made ofthe BMRR epithermal neutron beam, starting each time with fission neutrons from the reactor core and following the neutrons or gamma rays produced by neutron capture to the patient irradiation port, to predict improvements of the beam parameters in air at the patient irradiation port. First, an MCNP calculation was made of the reactor and beam as it existed in May 1992 and benchmarked to measurements. Then, after each change, the calculated results at X and Y (Figures 2a and 2b) were compared to the May 1992 calculation. Each calculation was run for 12 hours on the mainframe at BNL. More than 300,000 neutrons sampled from the Watt fission spectrum in the fuel elements were tracked in each run to provide good statistics. The statistical error of these calculations is 4% for the epithermal neutron flux tallied in cells located at X and Y at the patient irradiation port. Separate MCNP runs with a neutron energy cutoff at 10 keV, which means neutrons are no longer tracked once they are slowed down to an energy less than 10 keV, were made to provide a 4% statistical error for the calculation of the fast neutron dose.
The Monte Carlo code Monte Carlo calculations are an appropriate technique to describe different fluxes and doses in a mixed radiation field. They obtain answers by simulating individual particles and recording some aspects of the average behavior of many such particles. Whether an individual particle (neutron or photon) will interact with a material and what kind of interaction will occur is determined based on rules (physics) and probabilities (transport data). The possibility distributions are randomly sampled from the transport data (nuclear data libraries) to determine the outcome at each step of the particle’s life. A particle starts from the source and continues along a path determined by the probability of reactions until its loss by absorption, con-
Measurements were made using bare and Cd-covered gold foils in air at the Bi face and in an ellipsoidal head phantom. The phantom was made by gluing 1.27 cmthick Lucite plates together and then shaping the stack of plates using a computer-controlled mill. The semi-axes of the phantom were 7.5, 7.5, and 9.8 cm. The volume of the phantom is the same as the head model used by Zamenhof (14) for treatment planning in BNCT. Lucite rods, which have slits at different depths (2, 4, 6, 8 cm, etc.) to accommodate bare and Cd-covered gold foils, can be inserted into the phantom for neutron flux measurements.
Boron neutron
capture
The thermal neutron flux at each depth ofthe ellipsoidal head phantom was determined by measurements of the induced gamma activity in bare and Cd-covered gold foils. The gold foils were 0.00 127-cm thick with an average mass of 9 mg and an average diameter of 0.7 cm. Gold has an epithermal neutron resonance at an energy of 4.9 eV. The effect of this resonance peak in measurements of the thermal neutron flux can be determined by the Cd difference method. First, a bare gold foil is used to measure the total activation from thermal and epithermal neutrons. Then a gold foil with a 0.1 cm-thick Cd cover to block thermal neutrons is used to measure the activation from epithermal neutrons only. The subtraction of the activity of the Cd-covered gold foil from the activity of the bare gold foil gives the activation from thermal neutrons alone (1 1). The gold foil activation method has been routinely used for neutron flux measurements at the BMRR and the measurements are consistent with MCNP calculations. Paired ionization chambers were used to measure the fast neutron and gamma dose rates in air at the irradiation port. First, a tissue equivalent (TE) ionization chamber filled with TE gas was used to measure the total absorbed dose in a mixed neutron/gamma field. The next step was to use a neutron-insensitive graphite wall chamber filled with CO? gas to measure the gamma dose component. Thus, the fast neutron dose was determined by subtracting the gamma dose component from the total absorbed dose (11). MCNP calculations were used to predict the outcome of a change of the design. With a promising prediction, this change was made to enhance the beam. Then, measurements of flux and dose were made to substantiate the improvements.
RESULTS Changes to upgrade the epithermal neutron beam may be possible in the reactor core. moderator tanks in the A, B, and C regions, and the patient irradiation port. Comparisons of the MCNP calculations and dosimetric measurements are discussed in the following sections.
therapy
1153
0 H. B. LIU et ul.
Table 2. Comparisons of thermal activation of Au foils at different depths in the ellipsoidal head phantom Thermal activity (Counts/minmgMW) Depth (em)
May 1992
June 1992
2
1.13 x 10’
4 6 8
1.11 x 10’ 8.12 x 106 5.24 X lo6
1.28 1.35 9.64 6.06
x x X x
10’ 10’ IO6 IOh
+13 +22 +19 +16
The epithermal neutron flux at the patient irradiation port may be increased by shifting the fuel elements (or power distribution) in the reactor core toward the patient irradiation port while retaining the reactor in a critical but controlled condition. Figure 3 shows a horizontal section of the BMRR core lattice. Each of the 32 empty squares represents the location of a fuel element, labelled as rows A through F. The four black bars represent the control rods. The A2 and A3 positions were not fueled before May 1992. The former was filled with a graphite spacer block, and A3 was a tube for sample irradiation. Space not fueled nor occupied by control rods and instrumentation within the Al vessel was filled with graphite or Al blocks. In June 1992. the BMRR was operated with one additional fuel element in place of the graphite spacer block in the A2 location. With the fuel element in position, measurements were made of flux and dose at the epithermal port to compare to previous values and MCNP calculations. A relative power level for this fuel element is assumed in the MCNP calculation, based on the power distribution data. The MCNP calculation predicted an 18% increase of the epithermal neutron flux and fast neutron dose rate with this fuel element: the measurements confirmed this prediction. The thermal neutron flux, determined by using bare and Cd-covered gold foil activation, at different depths up to 8 cm into the ellipsoidal head phantom went up (18 & 4)% on average, while the fast neutron and gamma dose rates. determined by using paired ionization chambers, went up no more than 18%. These comparisons are shown in Table 2 and Table 3. On the other side of the reactor, at the BMRR thermal port, the thermal neutron flux fell by 11%. Another calculation considered moving two fuel elements from the thermal side (F2 and F3) to the A2 and
Table 3. Comparisons
of fast neutron and gamma in air at the irradiation port
Fig. 3. Cross section of the BMRR core lattice labelled as rows A through F.
Difference %
Dose rate (cGy/minMW)
May 1992
Fast neutron Gamma
1.71 0.35
June
1992
2.00 0.40
dose rates
Difference % +17 +14
1. J. Radiation
1154
Oncology
0 Biology 0 Physics
4. Comparisons of thermal activation of Au foils at different depths in the ellipsoidal head phantom
Table
Thermal activity (Counts/minmgMW) Depth (cm) 2 4 6 8
Mav 1992 1.13 1.11 8.12 5.24
x x x X
IO7 10’ lo6 lo6
Ott 1992 1.65 1.68 1.24 7.93
X X X x
10’ IO7 10’ 106
Difference % f46 f51 +53 +51
A3 locations on the epithermal side. The results showed that the epithermal neutron flux would be increased by 36% at X; no significant change in the neutron spectrum change was predicted. The fast neutron dose per epitherma1 neutron increased no faster than did the epithermal neutron flux. In October 1992, the BMRR core was reconfigured. Four new fuel elements were placed in the first row (A 1, A2, A3, and A4) facing the epithermal port. Because of excessive reactivity in the reactor core, position B6 was filled with a graphite spacer block to bring the shutdown margin within the technical specifications. To minimize the effect imposed on the thermal neutron beam, the fuel elements on the thermal side of the reactor core were not changed. Dosimetric measurements were then made and compared to the values in May 1992. The measurements were repeated with reliable reproducibility: dilferences of less than 3% were recorded in data taken in October 1992 and January 1993. An increase of the epithermal neutron flux of more than 36% was expected because four new fuel elements replaced the four partially burned-up ones in the A row. The new measurements show that the thermal neutron flux at different depths up to 8 cm into the ellipsoidal head phantom went up (50 +3)% on average while the fast neutron and gamma dose rates went up only 36% and 34%, respectively (Tables 4 and 5). In addition to physical dosimeters, biological dosimeters are used to measure the epithermal neutron beam to give information about the doses and the RBEs ofthe different types of radiation in the beam. Chinese hamster cells (V79) were irradiated in air at the port face. Based on 10% cell survival as the endpoint and comparisons with measurements made before May 1992, the RBE for the fast neutrons in the beam in air has not changed and is 3.8 + 0.3. Since the physical doses are measured to determined this RBE and the RBE does not change, this finding also verifies that there has been a slight reduction of the fast neutron dose per epithermal neutron as indicated by the paired ionization chamber measurements. Shifting of the core while increasing the epithermal neutron flux will reduce the thermal neutron flux in the beam on the other side of the reactor. After this change, the thermal neutron flux at the thermal port fell by 22% compared to values before May 1992. Since the intensity of the BMRR thermal neutron beam is more than ade-
Volume
28, Number
5, 1994
quate for small animal and cell-culture irradiations, such a reduction in the thermal beam is acceptable. Table 1 shows the present beam parameters of the BMRR epithermal neutron beam at the irradiation port compared to the other two operating epithermal neutron beams available. These measurements verify the MCNP calculations which predicted that most of the epithermal neutrons reaching the patient irradiation port come from fission neutrons produced in the A and B rows of fuel elements on the epithermal side of the reactor core. By rearranging the fuel elements in the reactor core while maintaining criticality and control, the beam intensity at the epithermal patient irradiation port is effectively increased and its quality is slightly improved.
Possible changes in the moderator To redesign the moderator, the two empty spaces in tank C can be filled with Al pellets to move the moderator toward the reactor core. Then, some space is left in the beam shutter to accommodate the outer Bi shield which will allow an air indentation at the patient irradiation port to shape the emerging neutron beam toward the forward direction. The packing density of Al pellets was assumed to be 60% in the MCNP calculations. The present Al and Al2O3 plates are intermixed in the moderator tanks of the A and B regions. The idea for improving the beam parameters at the patient irradiation port was to change the moderator’s configuration and the amount of Al and A1203. Calculated fluxes for a change to a block of Al followed by a block of A&O3 in the A and B tanks, with the same total thickness of Al and A1203 as in the present moderator tanks, were compared to the calculated fluxes of the present moderator design. With the same epithermal neutron flux at the patient irradiation port. the fast neutron dose per epithermal neutron is reduced by 30%, significantly improving beam quality. The outer Bi block shields the patient from neutroninduced gamma rays from the construction material and Al in the moderator tanks. Another alternative is to use Pb plus 0.05% atomic number density of 6Li to replace Bi. The ‘Li captures the thermal neutrons while the Pb is a relatively effective moderator of fast neutrons compared to the Bi. Calculations indicated that a change from Bi to Pb(Li) would slightly reduce the fast neutron dose per epithermal neutron. The Li-poly assembled around the Bi port slightly increases the background gamma dose due to ‘H-induced
Table 5. Comparisons of fast neutron and gamma in air at the irradiation port
dose rates
Dose rate (cGv/minMW)
May 1992
Difference %
Fast neutron Gamma
1.71 0.35
June
1992
2.33 0.47
$36 +34
1155
Boron neutron capture therapy 0 H. B. LIU et al.
equal to 0.57 at X. With about 8 cm shift from X to Y, the beam directionality becomes 0.67, but with a loss of about 35% of the epithermal neutron flux. If these proposed changes to the beam moderator and the irradiation port are incorporated into the present configuration of the reactor core, the design calculations indicate that the epithermal neutron flux can be increased by 100% at X and I’, while the fast neutron dose per epithermal neutron can be reduced by 38% at X and Y compared to beam parameters in May 1992; the thermal neutron flux and gamma dose per epithermal neutron will remain at low levels. Table 6 shows the relative change in beam parameters at the BMRR epithermal port. With the proposed design, the calculated epithermal neutron flux at X is 3.6 X lo9 n/cm2sec at 3 MW power, while the fast neutron dose per epithermal neutron is 3.0 X lo-” cGycm2/n. For a patient at Y, the epithermal neutron flux will be 2.4 X lo9 n/cm2sec at 3 MW while the fast neutron dose per epithermal neutron still will be 3.0 X lo-” cGycm2/n. Accordingly, such an improvement can reduce an irradiation treatment to 27 minutes under conditions where 10 ppm of “B is uniformly distributed in the head and the center of the brain reaches its tolerance dose limit of 10 RBE-Gy. Also, the skin RBE dose can be reduced by 19%. The directionality of the beam can be increased from 0.67 to 0.72 at Y so that it penetrates deeper into the head.
Li-Poly Shield
I Beam Shutter
Fig. 4. Proposed facility.
design of the BMRR epithermal
Pb
I IT neutron
Li-Poly Shield
beam
gamma rays from the Li-poly shield; one solution is to replace the Li-poly material with pure Li2C03 sheets. The gamma dose per epithermal neutron is reduced by 10% at X. This change is not considered to be worthwhile since without H in the shield, the pure Li2C03 sheets will not be as effective as the Li-poly shield in moderating the fast neutron flux outside the port. Proposed moduutor and irradiation port designs The final design of the BMRR with the proposed reconfiguration in the reactor core and the redesign of the moderator and the patient irradiation port is depicted in Figure 4. The major changes, including the October 1992 core configuration, include: (a) the C tank will be filled with Al pellets: (b) a moderator will be made with a combined optimized thickness of 22 cm of Al followed by 41 cm of A1203 in the A, B, and C regions; (c) a 13 cm-thick Pb shield with 0.05% atomic number density of ‘Li will be the gamma shield following the proposed moderator; and (d) a 16 cm-long air indentation will be created at the patient irradiation port from the face of the Pb(Li) shield to the Y point. The present epithermal neutron beam at the irradiation port is nearly isotropic with a neutron current-to-flux ratio (i.e.. beam directionality)
Table 6. Relative change in beam parameters
Old beam at X (May 1992) Old beam at Y (May 1992) Present beam at X (Ott 1992)* Present beam at Y (Ott 1992)* Beam with proposed moderator at Xt Beam with proposed moderator at Yt * Values based on MCNP calculations ’ Values based on MCNP calculations.
CONCLUSION Improvements for the BMRR epithermal neutron beam were evaluated by MCNP calculations. The proposed changes to enhance the beam parameters at the patient irradiation port included rearranging the fuel elements in the reactor core and redesigning the beam moderator and the patient irradiation port. The MCNP calculated values for the previous and proposed designs were compared to demonstrate that these changes would improve the beam. The results show that the epithermal neutron flux can be increased by 100% and the fast neutron dose per epitherma1 neutron can be reduced by 38% while the beam directionality can be improved from 0.67 to 0.72. At the same time, the thermal neutron flux and beam gamma
at 3 MW power at the BMRR epithermal
Epithermal neutron flux IO9 (n/cm2 set)
Fast neutron dose per epi-neutron 10-l ’ cGycm2/n
1.8 1.2 2.7 1.8 3.6 2.4
4.8 4.8 4.3 4.3 3.0 3.0
and verified by measurements.
port Gamma dose per epi-neutron lo-” cGycm*/n 1.1 1.4 1.0 1.3 1.1 1.4
II56
I. J. Radiation
Oncology
0 Biology
0 Physics
dose per epithermal neutron will stay at the same level. With these improvements, a BNCT treatment can be completed in 27 minutes, with a 19% reduction in skin RBE dose. When the proposed reconfiguration of the core was carried out, dosimetric measurements confirmed the
Volume 2X. Number 5. 1094
MCNP calculations. The present epithermal neutron beam at the BMRR gained a 50%’ increase in the epithermal neutron flux and the fast neutron and gamma doses per epithermal neutron were reduced by 9% and I I%, respectively, after the recent (October 1992) rearrangement of the fuel elements in the reactor core.
REFERENCES I. Asbury. A. K.: Ojemann. R. G.: Nielsen. S. L.: Sweet. W. H. Neuropathologic study of fourteen cases of malignant brain tumor treated by Boron-10 slow neutron capture radiation. J. Neuropathol. Exp. Neural. 31:278-303: 1972. 2. Barth. R. F.; Soloway, A. H.: Fairchild, R. G. Boron neutron capture therapy for cancer. Sci. Am. Oct. 100-107: 1990. 3. Becker, G. K.: Harker. Y. D.: Miller. L. G.: Andre]. R. A.; Wheeler, F. J. Neutron spectrum measurements in the aluminum oxide filtered beam facility at the Brookhaven Medical Research Reactor. In: Harling, 0. K.. Bernard. J. A.. Zamenhof, R., eds. Neutron beam design. development, and performance for neutron capture therapy. New York: Plenum Press; 1990:235-245. 4. Briesmeister. J. F., ed. MCNP-A general Monte Carlo code for neutron and photon transport. Los Alamos National Laboratory. LA-7396-M. Rev. 2; 1986. 5. Brugger, R. M.: Less. T. J.: Passmore. G. G. An intermediateenergy neutron beam for NCT at MURR. BNL-51994. Brookhaven National Laboratory: 1986. 6. Choi. J. R.; Zamenhof, R. G.: Yanch, J. C.: Rogus, R.: Harling, 0. K. Performance of the currently available epithermal neutron beam at the Massachusetts Institute of Technology Reactor (MITR-II). In: Allen, B. J.. Moore, D. E., Harrington. B. V.. cds. Progress in neutron capture therapy for cancer. New York: Plenum Press: 199253-56. 7. Fairchild, R. G.: Kalef-Ezra. J.: Saraf, S. K.: Fiarman, S.: Ramsay, E.; Wielopolski. L.; Laster. B. H.: Wheeler, F. J. Installation and testing of an optimized epithermal neutron beam at the Brookhaven Medical Research Reactor (BMRR). In: Harling, 0. K.. Bernard. J. A.. Zamenhof. R.. eds. Neutron beam design, development, and performance for neutron capture therapy. New York: Plenum Press; 1990: 185-199. 8. Farr. L. E.: Sweet, W. H.: Robertson. J. S.: Foster, C. G.:
9.
IO.
I I.
12.
13.
14.
Locksley. H. B.; Sutherland. D. L.: Mendelsohn. M. L.: Stickle\;. E. E. Neutron capture therapy with boron in the treatment of glioblastoma multiforme. Am. J. Roentgenol. 7 1:279-293; 1954. Godel, J. B. Description of facilities and mechanical components. Medical Research Reactor (MRR). BNL 600 (T173), Brookhaven National Laboratory; Feb. 1960. Liu. H. B. Isodose curves and treatment planning for boron neutron capture therapy. Dissertation in Nuclear Engineering Program of University of Missouri, Columbia: Aug. 1992. Saraf. S. K.; Kalef-Ezra. J.: Fairchild, R. G.: Laster, B. H.; Fiarman. S.: Ramsay. E. Epithermal beam development at the BMRR: Dosimetric evaluation. In: Harling, 0. K., Bernard, J. A.. Zamenhof, R., eds. Neutron beam design, development. and performance for neutron capture therapy. New York: Plenum Press: I990:307-3 16. Slatkin. D. N. A history of boron neutron capture therapy of brain tumors-Postulation of a brain radiation dose tolerance limit. Brain Il4:1609-1629: 1991. Watkins, P.: Constantine, G.: Stecher-Rasmussen, F.: Freudenreich. W.: Moss. R. L.: Ricchena. R. MCNP calculations for the design and characterization ofthe Petten BNCT epithermal neutron beam. In: Allen, B. J., Moore. D. E.. Harrington. B. V., eds. Progress in neutron capture therapy for cancer. New York: Plenum Press: 1992:71-73. Zamenhof. R.; Brenner. J.; Yanch. J.: Wazer, D.: MadocJones. H.: Saris, S.; Harling, 0. Treatment planning for neutron capture therapy of glioblastoma multiforme using an epithermal neutron beam from the MITR-II research reactor and Monte Carlo simulation. In: Allen B. J., Moore. D. E.. Harrington. B. V.. eds. Progress in neutron capture therapy for cancer. New York: Plenum Press: 1992: l73177.
Oncology
Pergamon
Biol. Phys., Vol. 28, No. 5, pp. 1149-I 156, 1994 Copyright 0 1994 Elsevier Science Ltd Printed in the USA. All rights reserved 0360-3016/94 $6.00 + .OO
0360-3016(94)E0076-V
??Special Feature- Physics Original Contribution
ENHANCEMENT OF THE EPITHERMAL NEUTRON BEAM USED FOR BORON NEUTRON CAPTURE THERAPY HUNGYUAN DAVID
B. Lru,
C.
RORER,
*Medical
PH.D.,*
Department
M. BRUGGER,
ROBERT
PH.D.,?
JIH-PERNG
and tReactor
PH.D.,*
DENNIS D. GREENBERG, B.S.,*
Hu, PH.D.-~ AND HENRY M. HAUPTMAN,
Division,
Brookhaven
National
Laboratory,
Upton,
B.S.-j-
NY 11973
Purpose: This report describes a study to enhance the epithermal neutron beam at the Brookhaven Medical Research Reactor by increasing the epithermal neutron flux and/or reducing contamination by fast neutrons. Methods and Materials: The beam was reevaluated using Monte Carlo calculations and flux and dose measurements in air and in an ellipsoidal head phantom at the patient irradiation port. Changes in its geometry and materials were considered, including rearranging the fuel elements in the reactor core and redesigning the moderator and the patient irradiation port. Results: Calculations of the new fluxes and doses at the patient irradiation port showed that the epithermal neutron flux can be increased by lOO%, while the fast neutron dose per epithermal neutron can be reduced by 38%. In 1992, some of the proposed changes were made. In June 1992, measurements were made after one additional fuel element was added to replace a graphite spacer block on the epithermal beam side of the reactor core. The results show that the epithermal neutron flux increased by 18%, as predicted by the Monte Carlo calculations. In October 1992, the fuel elements in the reactor core were rearranged by placing four new fuel elements in the first row facing the epithermal port; the intensity of the epithermal neutron beam increased by 50% and the fast neutron and gamma doses per epithermal neutron may have decreased slightly. Conclusion: The epithermal neutron beam at the Brookhaven Medical Research Reactor has gained a 50% increase in the epithermal neutron flux and the fast neutron and gamma doses per epithermal neutron are reduced slightly after the rearrangement of the fuel elements in the core. Boron neutron capture therapy, Brookhaven dosimetry.
Medical Research
Epithermal
neutron beam, Neutron
lignant tissue while the dose imposed on the healthy tissue is below the threshold for irreversible damage. The initial clinical trials of BNCT in the 1950s at the Brookhaven National Laboratory (BNL) were made with a beam of thermal neutrons from the Brookhaven Graphite Research Reactor (BGRR) (8). The Brookhaven Medical Research Reactor (BMRR) (9) was built and became operational in 1959 to provide a higher flux beam of thermal neutrons for BNCT research. After the initial trials, it was realized that to reach deep-seated brain tumors, a beam of epithermal neutrons was preferable to a beam of thermal neutrons. The depth of penetration of a beam of thermal neutrons is limited, so excessive damage to the skin and crania1 structures occurs. In contrast, epithermal
INTRODUCTION
Optimal cancer therapy would eliminate tumor cells without seriously damaging the normal tissue. Today’s standard treatments have succeeded in curing some kinds of cancers, but for others, such as malignant brain tumors, we have achieved only marginal lengthening of the patient’s life. New treatment modalities are needed to selectively irradiate the tumor cells while sparing the normal tissue. Boron neutron capture therapy (BNCT) (2) is based on a two-part modality to treat these resistant tumors. It brings together two components, namely l”B nuclei and thermal neutrons, that, when used separately, have only minor effects on the normal cells. The combination of these two components at the tumor site releases intense radiation of (Y and Li particles that can destroy the ma-
Data presented, in part, at the Fifth International
Reactor (BMRR),
neutrons, keV,
Symposium
have
those
having
a greater
energies
depth
between
of penetration
0.4 eV and so that
10
a deep-
We thank C. R. Gordon and L. S. Warkentien for their handling of the cell survival measurements and A. L. Ruggiero for clerical assistance. We also thank A. D. Woodhead for final editing of this manuscript. This research was supported by the US Department of Energy under contract DE-AC02-76CHOOO 16 with Brookhaven National Laboratory. Accepted for publication 2 September 1993.
on Neutron Capture Therapy, Columbus, OH, 13- 17 September 1992. Reprint requests to: Hungyuan B. Liu, Ph.D., Medical Dept., Building 490, Brookhaven National Laboratory, Upton, NY 11973. Acknowledgement-We thank J. Gajewski for assistance preparing the geometrical design for the Monte Carlo calculations. 1149
1150
1. J. Radiation Oncology 0 Biology 0 Physics
seated tumor can be treated while sparing surface tissue. In 1988, a moderator was inserted into one of the beam shutters at the BMRR to produce a beam of epithermal neutrons. This beam, which has a relatively low contamination of fast neutrons. has been shown to be satisfactory for BNCT. A patient can be fully treated in about an hour. As a practical matter, the ability of patients to tolerate the psychological stress associated with lying motionless in a neutron beam for an hour is a concern. Therefore, to shorten the treatment time, it is worthwhile to increase the epithermal neutron flux. Also, while the present BMRR epithermal neutron beam is satisfactory for BNCT, it would be more effective if the fast neutron dose could be reduced to spare the skin more. The purpose of this study was to show how the epithermal neutron beam at the BMRR can be improved by increasing the flux of epithermal neutrons and reducing the fraction of fast neutrons. We used Monte Carlo calculations to predict the changes in fluxes produced by changes in the material or geometry in the core of the BMRR and in the moderator of the beam shutter and the patient irradiation port. Several of the proposed changes were made in 1992 and new measurements are compared to earlier ones.
BACKGROUND
OF
BNCT
AT BNL
Eurly triuls qf BNCT Initial clinical trials of BNCT began in 195 1 using a thermal neutron beam at the BGRR. By removing part of the concrete shielding over the top of the BGRR, an irradiation port was built to deliver thermal neutrons to irradiate brain tumors in patients. Immediately after the intravenous injection of a borax solution, the patient’s head was fixed in the proper position over the irradiation port. The reactor was then brought to full power to irradiate the patient for about half an hour after which the reactor was shut down. The patients were irradiated by thermal neutrons with the inevitable contamination by fast neutrons and gamma rays. During the 195Os, 28 patients were treated at the BGRR. The BMRR became operational in 1959 to deliver a more intense beam of thermal neutrons for BNCT, with an improved beam-extraction facility that facilitated future beam development. Between 1959 and 1961, 17 patients were treated at the BMRR. Analysis of the results of these early clinical trials showed they had little significant therapeutic effect (12). One patient in this series of BNCT trials did not deteriorate neurologically after treatment; however, he died 5 months after treatment from widespread extracranial metastases. At necropsy, there was no evidence of viable tumor tissue at the primary site. Concurrently, efforts toward developing BNCT were also made at the Massachusetts Institute of Technology Reactor (MITR). The outcome of the MIT clinical trials of BNCT was also unsatisfactory ( 1). Failure to demonstrate substantial extensions of lifespan resulted in suspensions of both the BNL and the MIT clinical trials of BNCT.
Volume
28, Number
5, 1994
No patients have been treated by BNCT in the United States since 196 1. The experimenters concluded that the ‘“B compound did not concentrate and persist in the tumor cells, and that the thermal neutrons did not penetrate to the adequate depths. Therefore, there was excessive damage to the normal tissue and inadequate destruction of the deep-seated brain tumor. Better “B compounds that target tumors were needed together with beams of epithermal neutrons that would penetrate to deep-seated tumors. BMRR The BMRR is an adjunct facility to the BNL’s Medical Research Center, and is located in an 18.3 m-diameter gas-tight confinement building which is connected to the laboratory through an air lock. Figure 1 shows the cutaway configuration of the BMRR. There are two treatment neutron beam ports, each within a shielded room, on opposite faces of the reactor. At present, one is a thermal neutron beam and the other is an epithermal neutron beam. The emerging neutron beams at the patient irradiation ports are interrupted by beam shutters with assemblies that can be raised and lowered hydraulically inside a vertical cavity to control the irradiation. When the shutter is down, a high-density concrete section of each shutter blocks the beam between the reactor core and the patient location. When the shutter is raised, the moderator section of the shutter is between the reactor core and the patient location. When the moderator is D20, a thermal neutron beam is produced and when the moderator is Al and A1203, an epithermal neutron beam is produced.
Fig. 1. Cut-away view of the BMRR, showing (A) the reactor core, (B) the beam shutter, and (C) the moderator in the beam shutter.
Boron neutron capture therapy 0 H. B.
II
within the Al vessel is filled with graphite or Al blocks. The control system consists of three safety rods and one regulating rod. An air-cooled reflector structure of graphite blocks surrounds the reactor core. Two stationary Bi block walls in the beam’s direction within the reflector were added to shield patients from gamma rays from the reactor core. Figure 2a depicts a horizontal section of the central plane of the BMRR extending from the reactor core out to the epithermal patient irradiation port, showing the reactor core, graphite reflector, inner Bi shield, moderator tanks in the A, B, and C regions, beam shutter, outer Bi shield, and Li2C03-in-polyethylene (Li-poly) shield at the patient irradiation port. The irradiation points at the irradiation port are labeled X and Y. In 1959, the reactor started up with 17 new fuel elements; with time. the fuel was burned up, but the partially burned elements were not removed and new elements were added until May 1992 when there were 30 partially spent fuel elements in the reactor core. In June 1992, one extra fuel element was added to the reactor core in place of a graphite spacer block. The reactor core was reconfigured in October 1992 and there are now 3 1 fuel elements in the reactor core with four new fuel elements facing the epithermal irradiation port. The moderator region C consists of two empty spaces with areas of 127 X 119 cm2 and 4 and 8 cm thick, respectively. These empty spaces can be filled with liquid or pellets. The shutters are 1.9 cm-thick steel shells, filled with high density concrete at the top and the moderator tanks A and B near the bottom. The Al and A1203, which were identified by Brugger and Less (5) as good moderators to produce an epithermal neutron beam, were selected by Fairchild and Wheeler (7) as the primary moderator and used to fill in the A and B regions of the existing beam shutter to produce the present epithermal neutron beam. Figure 2b shows the A, B, and C regions of the present neutron moderator. The moderator is arranged in an alternate sequence of Al and AlzOj, followed by the Bi shield both in the beam shutter and at the patient irradiation port. A 3.8 cm-thick Li-poly shield was added in 199 1 around the Bi port to reduce the neutron flux coming from outside the port.
High Density Li-Poly Shield
Gmcrcte
J -Bi
D
1Pbl +
II
a
Li-Poly
I I I Shic’d
Beam’Sbutter
Li-Poly Shield
LiPoly Shield
b Fig. 2. (a) Horizontal section of the present BMRR epithermal neutron beam, showing the moderator tanks in the A, B, and C regions. (b) Horizontal section of the present BMRR epithermal neutron beam, showing the mixing assembly of Al and A&O3 in moderator tanks A and B.
The reactor core is cooled by an upward flow of water and operates at a maximum power of 3 MW. Most components of the reactor core, including the reactor vessel and cooling water system, are made of Al. Space not occupied by fuel elements, control rods, or instrumentation
Table
BMRR at BMRR at BMRR at BMRR at MITR* PETTEN+
X Y X Y
(May (May (Ott (Ott
1992) 1992) 1992) 1992)
* Personal communication, MA, USA. + Personal communication,
1. Epithermal
neutron
Power (MW
Epithermal neutron flux lo9 (n/cm* set)
3 3 3 3 5 45
1.8 1.2 2.7 1.8 0.20 0.33
Dr. 0. Harling,
Nuclear Engineering
Dr. R. Moss, Commission
1151
LIU et a/
beams for BNCT Fast neutron dose per epi-neutron lo-” cGycm*/n
Gamma dose per epi-neutron 10-i ’cGycm*/n
4.8 4.8 4.3 4.3 13 10.4 Department,
ofthe European
Massachusetts
Communities,
1.1 1.4 1.0 1.3 14 8.4 Institute
of Technology,
Cambridge,
Joint Research Centre, Petten, The Netherlands.
1152
I. J. Radiation Oncology 0 Biology 0 Physics
Since the beam moderator was assembled in 1988, dosimetric measurements have been carried out in air and in phantoms at the epithermal irradiation port (3, 10, 11) to define the incident flux and the imparted dose as a function of energy in the epithermal neutron beam. Compared to the other two operating epithermal neutron beams, one in Cambridge, MA and one in Petten, The Netherlands, (6,13) the BMRR epithermal neutron beam delivers the highest epithermal neutron flux and the lowest fast neutron dose per epithermal neutron. Table 1 compares the parameters of these epithermal neutron beams. Liu (10) calculated treatment plans for treating brain tumors by irradiation from the BMRR epithermal neutron beam. In one such calculation, the patient is placed at Y (Figures 2a and 2b), and is irradiated with the BMRR epithermal neutron beam at 3 MW power, with a 10 ppm “B uniformly distributed through the head as a baseline concentration, an irradiation of 54 min is required to reach a dose tolerance limit of 10 RBE-Gy delivered to the center of the brain ( 12). The RBE-Gy stands for the relative biological effectiveness (RBE) of a radiation component multiplied by the physical dose of the component (Gy). Compared to traditional external beam radiation therapy which usually is completed in a few minutes per fractionation, a reduction of the 54-min irradiation time at the BMRR becomes important to ameliorate the patient’s anxiety. In addition, a reduction of the fast neutrons in the beam is also important to reduce the dose to the patient’s skin, since the fast neutron RBE dose contributes over 50% of the dose to the skin.
METHODS
AND
MATERIALS
To optimize a physical facility, a reliable calculation technique is needed to predict the outcome of a change of the design, with dependable experimental methods to validate this change. In this study of upgrading the BMRR epithermal neutron beam, a Monte Carlo code was used to predict the improvements, and then, where feasible, dosimetric measurements in air and in an ellipsoidal head phantom were made and repeated to substantiate the Monte Carlo predictions.
Volume 28, Number 5, 1994
version, or escape. Then, another particle is sampled to start the interaction with materials all over again and so on, for thousands to millions of particles. The Monte Carlo for Neutron and Photon (MCNP) transport code, which was developed at the Los Alamos National Laboratory (4) is a general-purpose Monte Carlo code, which can be used to model the transport of neutrons, photons, coupled neutrons/photons, or coupled photons/electrons (MCNP version 4.2). The code treats an arbitrary 3-D configuration of materials in geometric cells bounded by planes, spheres, cones, or ellipsoids. MCNP uses either continuous or discrete nuclear crosssection libraries to describe the probability of reactions. Nuclear data is tabulated for neutron interactions, photon interactions, thermal particle scattering, and neutron dosimetry and activation. The code is well designed and experimentally validated, so it is a good computational tool for applications in the neutron beam designs of BNCT. A detailed MCNP input program was designed to simulate different areas of interest for the BMRR core and the epithermal neutron beam facility. MCNP (version 4.0) calculations were made ofthe BMRR epithermal neutron beam, starting each time with fission neutrons from the reactor core and following the neutrons or gamma rays produced by neutron capture to the patient irradiation port, to predict improvements of the beam parameters in air at the patient irradiation port. First, an MCNP calculation was made of the reactor and beam as it existed in May 1992 and benchmarked to measurements. Then, after each change, the calculated results at X and Y (Figures 2a and 2b) were compared to the May 1992 calculation. Each calculation was run for 12 hours on the mainframe at BNL. More than 300,000 neutrons sampled from the Watt fission spectrum in the fuel elements were tracked in each run to provide good statistics. The statistical error of these calculations is 4% for the epithermal neutron flux tallied in cells located at X and Y at the patient irradiation port. Separate MCNP runs with a neutron energy cutoff at 10 keV, which means neutrons are no longer tracked once they are slowed down to an energy less than 10 keV, were made to provide a 4% statistical error for the calculation of the fast neutron dose.
The Monte Carlo code Monte Carlo calculations are an appropriate technique to describe different fluxes and doses in a mixed radiation field. They obtain answers by simulating individual particles and recording some aspects of the average behavior of many such particles. Whether an individual particle (neutron or photon) will interact with a material and what kind of interaction will occur is determined based on rules (physics) and probabilities (transport data). The possibility distributions are randomly sampled from the transport data (nuclear data libraries) to determine the outcome at each step of the particle’s life. A particle starts from the source and continues along a path determined by the probability of reactions until its loss by absorption, con-
Measurements were made using bare and Cd-covered gold foils in air at the Bi face and in an ellipsoidal head phantom. The phantom was made by gluing 1.27 cmthick Lucite plates together and then shaping the stack of plates using a computer-controlled mill. The semi-axes of the phantom were 7.5, 7.5, and 9.8 cm. The volume of the phantom is the same as the head model used by Zamenhof (14) for treatment planning in BNCT. Lucite rods, which have slits at different depths (2, 4, 6, 8 cm, etc.) to accommodate bare and Cd-covered gold foils, can be inserted into the phantom for neutron flux measurements.
Boron neutron
capture
The thermal neutron flux at each depth ofthe ellipsoidal head phantom was determined by measurements of the induced gamma activity in bare and Cd-covered gold foils. The gold foils were 0.00 127-cm thick with an average mass of 9 mg and an average diameter of 0.7 cm. Gold has an epithermal neutron resonance at an energy of 4.9 eV. The effect of this resonance peak in measurements of the thermal neutron flux can be determined by the Cd difference method. First, a bare gold foil is used to measure the total activation from thermal and epithermal neutrons. Then a gold foil with a 0.1 cm-thick Cd cover to block thermal neutrons is used to measure the activation from epithermal neutrons only. The subtraction of the activity of the Cd-covered gold foil from the activity of the bare gold foil gives the activation from thermal neutrons alone (1 1). The gold foil activation method has been routinely used for neutron flux measurements at the BMRR and the measurements are consistent with MCNP calculations. Paired ionization chambers were used to measure the fast neutron and gamma dose rates in air at the irradiation port. First, a tissue equivalent (TE) ionization chamber filled with TE gas was used to measure the total absorbed dose in a mixed neutron/gamma field. The next step was to use a neutron-insensitive graphite wall chamber filled with CO? gas to measure the gamma dose component. Thus, the fast neutron dose was determined by subtracting the gamma dose component from the total absorbed dose (11). MCNP calculations were used to predict the outcome of a change of the design. With a promising prediction, this change was made to enhance the beam. Then, measurements of flux and dose were made to substantiate the improvements.
RESULTS Changes to upgrade the epithermal neutron beam may be possible in the reactor core. moderator tanks in the A, B, and C regions, and the patient irradiation port. Comparisons of the MCNP calculations and dosimetric measurements are discussed in the following sections.
therapy
1153
0 H. B. LIU et ul.
Table 2. Comparisons of thermal activation of Au foils at different depths in the ellipsoidal head phantom Thermal activity (Counts/minmgMW) Depth (em)
May 1992
June 1992
2
1.13 x 10’
4 6 8
1.11 x 10’ 8.12 x 106 5.24 X lo6
1.28 1.35 9.64 6.06
x x X x
10’ 10’ IO6 IOh
+13 +22 +19 +16
The epithermal neutron flux at the patient irradiation port may be increased by shifting the fuel elements (or power distribution) in the reactor core toward the patient irradiation port while retaining the reactor in a critical but controlled condition. Figure 3 shows a horizontal section of the BMRR core lattice. Each of the 32 empty squares represents the location of a fuel element, labelled as rows A through F. The four black bars represent the control rods. The A2 and A3 positions were not fueled before May 1992. The former was filled with a graphite spacer block, and A3 was a tube for sample irradiation. Space not fueled nor occupied by control rods and instrumentation within the Al vessel was filled with graphite or Al blocks. In June 1992. the BMRR was operated with one additional fuel element in place of the graphite spacer block in the A2 location. With the fuel element in position, measurements were made of flux and dose at the epithermal port to compare to previous values and MCNP calculations. A relative power level for this fuel element is assumed in the MCNP calculation, based on the power distribution data. The MCNP calculation predicted an 18% increase of the epithermal neutron flux and fast neutron dose rate with this fuel element: the measurements confirmed this prediction. The thermal neutron flux, determined by using bare and Cd-covered gold foil activation, at different depths up to 8 cm into the ellipsoidal head phantom went up (18 & 4)% on average, while the fast neutron and gamma dose rates. determined by using paired ionization chambers, went up no more than 18%. These comparisons are shown in Table 2 and Table 3. On the other side of the reactor, at the BMRR thermal port, the thermal neutron flux fell by 11%. Another calculation considered moving two fuel elements from the thermal side (F2 and F3) to the A2 and
Table 3. Comparisons
of fast neutron and gamma in air at the irradiation port
Fig. 3. Cross section of the BMRR core lattice labelled as rows A through F.
Difference %
Dose rate (cGy/minMW)
May 1992
Fast neutron Gamma
1.71 0.35
June
1992
2.00 0.40
dose rates
Difference % +17 +14
1. J. Radiation
1154
Oncology
0 Biology 0 Physics
4. Comparisons of thermal activation of Au foils at different depths in the ellipsoidal head phantom
Table
Thermal activity (Counts/minmgMW) Depth (cm) 2 4 6 8
Mav 1992 1.13 1.11 8.12 5.24
x x x X
IO7 10’ lo6 lo6
Ott 1992 1.65 1.68 1.24 7.93
X X X x
10’ IO7 10’ 106
Difference % f46 f51 +53 +51
A3 locations on the epithermal side. The results showed that the epithermal neutron flux would be increased by 36% at X; no significant change in the neutron spectrum change was predicted. The fast neutron dose per epitherma1 neutron increased no faster than did the epithermal neutron flux. In October 1992, the BMRR core was reconfigured. Four new fuel elements were placed in the first row (A 1, A2, A3, and A4) facing the epithermal port. Because of excessive reactivity in the reactor core, position B6 was filled with a graphite spacer block to bring the shutdown margin within the technical specifications. To minimize the effect imposed on the thermal neutron beam, the fuel elements on the thermal side of the reactor core were not changed. Dosimetric measurements were then made and compared to the values in May 1992. The measurements were repeated with reliable reproducibility: dilferences of less than 3% were recorded in data taken in October 1992 and January 1993. An increase of the epithermal neutron flux of more than 36% was expected because four new fuel elements replaced the four partially burned-up ones in the A row. The new measurements show that the thermal neutron flux at different depths up to 8 cm into the ellipsoidal head phantom went up (50 +3)% on average while the fast neutron and gamma dose rates went up only 36% and 34%, respectively (Tables 4 and 5). In addition to physical dosimeters, biological dosimeters are used to measure the epithermal neutron beam to give information about the doses and the RBEs ofthe different types of radiation in the beam. Chinese hamster cells (V79) were irradiated in air at the port face. Based on 10% cell survival as the endpoint and comparisons with measurements made before May 1992, the RBE for the fast neutrons in the beam in air has not changed and is 3.8 + 0.3. Since the physical doses are measured to determined this RBE and the RBE does not change, this finding also verifies that there has been a slight reduction of the fast neutron dose per epithermal neutron as indicated by the paired ionization chamber measurements. Shifting of the core while increasing the epithermal neutron flux will reduce the thermal neutron flux in the beam on the other side of the reactor. After this change, the thermal neutron flux at the thermal port fell by 22% compared to values before May 1992. Since the intensity of the BMRR thermal neutron beam is more than ade-
Volume
28, Number
5, 1994
quate for small animal and cell-culture irradiations, such a reduction in the thermal beam is acceptable. Table 1 shows the present beam parameters of the BMRR epithermal neutron beam at the irradiation port compared to the other two operating epithermal neutron beams available. These measurements verify the MCNP calculations which predicted that most of the epithermal neutrons reaching the patient irradiation port come from fission neutrons produced in the A and B rows of fuel elements on the epithermal side of the reactor core. By rearranging the fuel elements in the reactor core while maintaining criticality and control, the beam intensity at the epithermal patient irradiation port is effectively increased and its quality is slightly improved.
Possible changes in the moderator To redesign the moderator, the two empty spaces in tank C can be filled with Al pellets to move the moderator toward the reactor core. Then, some space is left in the beam shutter to accommodate the outer Bi shield which will allow an air indentation at the patient irradiation port to shape the emerging neutron beam toward the forward direction. The packing density of Al pellets was assumed to be 60% in the MCNP calculations. The present Al and Al2O3 plates are intermixed in the moderator tanks of the A and B regions. The idea for improving the beam parameters at the patient irradiation port was to change the moderator’s configuration and the amount of Al and A1203. Calculated fluxes for a change to a block of Al followed by a block of A&O3 in the A and B tanks, with the same total thickness of Al and A1203 as in the present moderator tanks, were compared to the calculated fluxes of the present moderator design. With the same epithermal neutron flux at the patient irradiation port. the fast neutron dose per epithermal neutron is reduced by 30%, significantly improving beam quality. The outer Bi block shields the patient from neutroninduced gamma rays from the construction material and Al in the moderator tanks. Another alternative is to use Pb plus 0.05% atomic number density of 6Li to replace Bi. The ‘Li captures the thermal neutrons while the Pb is a relatively effective moderator of fast neutrons compared to the Bi. Calculations indicated that a change from Bi to Pb(Li) would slightly reduce the fast neutron dose per epithermal neutron. The Li-poly assembled around the Bi port slightly increases the background gamma dose due to ‘H-induced
Table 5. Comparisons of fast neutron and gamma in air at the irradiation port
dose rates
Dose rate (cGv/minMW)
May 1992
Difference %
Fast neutron Gamma
1.71 0.35
June
1992
2.33 0.47
$36 +34
1155
Boron neutron capture therapy 0 H. B. LIU et al.
equal to 0.57 at X. With about 8 cm shift from X to Y, the beam directionality becomes 0.67, but with a loss of about 35% of the epithermal neutron flux. If these proposed changes to the beam moderator and the irradiation port are incorporated into the present configuration of the reactor core, the design calculations indicate that the epithermal neutron flux can be increased by 100% at X and I’, while the fast neutron dose per epithermal neutron can be reduced by 38% at X and Y compared to beam parameters in May 1992; the thermal neutron flux and gamma dose per epithermal neutron will remain at low levels. Table 6 shows the relative change in beam parameters at the BMRR epithermal port. With the proposed design, the calculated epithermal neutron flux at X is 3.6 X lo9 n/cm2sec at 3 MW power, while the fast neutron dose per epithermal neutron is 3.0 X lo-” cGycm2/n. For a patient at Y, the epithermal neutron flux will be 2.4 X lo9 n/cm2sec at 3 MW while the fast neutron dose per epithermal neutron still will be 3.0 X lo-” cGycm2/n. Accordingly, such an improvement can reduce an irradiation treatment to 27 minutes under conditions where 10 ppm of “B is uniformly distributed in the head and the center of the brain reaches its tolerance dose limit of 10 RBE-Gy. Also, the skin RBE dose can be reduced by 19%. The directionality of the beam can be increased from 0.67 to 0.72 at Y so that it penetrates deeper into the head.
Li-Poly Shield
I Beam Shutter
Fig. 4. Proposed facility.
design of the BMRR epithermal
Pb
I IT neutron
Li-Poly Shield
beam
gamma rays from the Li-poly shield; one solution is to replace the Li-poly material with pure Li2C03 sheets. The gamma dose per epithermal neutron is reduced by 10% at X. This change is not considered to be worthwhile since without H in the shield, the pure Li2C03 sheets will not be as effective as the Li-poly shield in moderating the fast neutron flux outside the port. Proposed moduutor and irradiation port designs The final design of the BMRR with the proposed reconfiguration in the reactor core and the redesign of the moderator and the patient irradiation port is depicted in Figure 4. The major changes, including the October 1992 core configuration, include: (a) the C tank will be filled with Al pellets: (b) a moderator will be made with a combined optimized thickness of 22 cm of Al followed by 41 cm of A1203 in the A, B, and C regions; (c) a 13 cm-thick Pb shield with 0.05% atomic number density of ‘Li will be the gamma shield following the proposed moderator; and (d) a 16 cm-long air indentation will be created at the patient irradiation port from the face of the Pb(Li) shield to the Y point. The present epithermal neutron beam at the irradiation port is nearly isotropic with a neutron current-to-flux ratio (i.e.. beam directionality)
Table 6. Relative change in beam parameters
Old beam at X (May 1992) Old beam at Y (May 1992) Present beam at X (Ott 1992)* Present beam at Y (Ott 1992)* Beam with proposed moderator at Xt Beam with proposed moderator at Yt * Values based on MCNP calculations ’ Values based on MCNP calculations.
CONCLUSION Improvements for the BMRR epithermal neutron beam were evaluated by MCNP calculations. The proposed changes to enhance the beam parameters at the patient irradiation port included rearranging the fuel elements in the reactor core and redesigning the beam moderator and the patient irradiation port. The MCNP calculated values for the previous and proposed designs were compared to demonstrate that these changes would improve the beam. The results show that the epithermal neutron flux can be increased by 100% and the fast neutron dose per epitherma1 neutron can be reduced by 38% while the beam directionality can be improved from 0.67 to 0.72. At the same time, the thermal neutron flux and beam gamma
at 3 MW power at the BMRR epithermal
Epithermal neutron flux IO9 (n/cm2 set)
Fast neutron dose per epi-neutron 10-l ’ cGycm2/n
1.8 1.2 2.7 1.8 3.6 2.4
4.8 4.8 4.3 4.3 3.0 3.0
and verified by measurements.
port Gamma dose per epi-neutron lo-” cGycm*/n 1.1 1.4 1.0 1.3 1.1 1.4
II56
I. J. Radiation
Oncology
0 Biology
0 Physics
dose per epithermal neutron will stay at the same level. With these improvements, a BNCT treatment can be completed in 27 minutes, with a 19% reduction in skin RBE dose. When the proposed reconfiguration of the core was carried out, dosimetric measurements confirmed the
Volume 2X. Number 5. 1094
MCNP calculations. The present epithermal neutron beam at the BMRR gained a 50%’ increase in the epithermal neutron flux and the fast neutron and gamma doses per epithermal neutron were reduced by 9% and I I%, respectively, after the recent (October 1992) rearrangement of the fuel elements in the reactor core.
REFERENCES I. Asbury. A. K.: Ojemann. R. G.: Nielsen. S. L.: Sweet. W. H. Neuropathologic study of fourteen cases of malignant brain tumor treated by Boron-10 slow neutron capture radiation. J. Neuropathol. Exp. Neural. 31:278-303: 1972. 2. Barth. R. F.; Soloway, A. H.: Fairchild, R. G. Boron neutron capture therapy for cancer. Sci. Am. Oct. 100-107: 1990. 3. Becker, G. K.: Harker. Y. D.: Miller. L. G.: Andre]. R. A.; Wheeler, F. J. Neutron spectrum measurements in the aluminum oxide filtered beam facility at the Brookhaven Medical Research Reactor. In: Harling, 0. K.. Bernard. J. A.. Zamenhof, R., eds. Neutron beam design. development, and performance for neutron capture therapy. New York: Plenum Press; 1990:235-245. 4. Briesmeister. J. F., ed. MCNP-A general Monte Carlo code for neutron and photon transport. Los Alamos National Laboratory. LA-7396-M. Rev. 2; 1986. 5. Brugger, R. M.: Less. T. J.: Passmore. G. G. An intermediateenergy neutron beam for NCT at MURR. BNL-51994. Brookhaven National Laboratory: 1986. 6. Choi. J. R.; Zamenhof, R. G.: Yanch, J. C.: Rogus, R.: Harling, 0. K. Performance of the currently available epithermal neutron beam at the Massachusetts Institute of Technology Reactor (MITR-II). In: Allen, B. J.. Moore, D. E., Harrington. B. V.. cds. Progress in neutron capture therapy for cancer. New York: Plenum Press: 199253-56. 7. Fairchild, R. G.: Kalef-Ezra. J.: Saraf, S. K.: Fiarman, S.: Ramsay, E.; Wielopolski. L.; Laster. B. H.: Wheeler, F. J. Installation and testing of an optimized epithermal neutron beam at the Brookhaven Medical Research Reactor (BMRR). In: Harling, 0. K.. Bernard. J. A.. Zamenhof. R.. eds. Neutron beam design, development, and performance for neutron capture therapy. New York: Plenum Press; 1990: 185-199. 8. Farr. L. E.: Sweet, W. H.: Robertson. J. S.: Foster, C. G.:
9.
IO.
I I.
12.
13.
14.
Locksley. H. B.; Sutherland. D. L.: Mendelsohn. M. L.: Stickle\;. E. E. Neutron capture therapy with boron in the treatment of glioblastoma multiforme. Am. J. Roentgenol. 7 1:279-293; 1954. Godel, J. B. Description of facilities and mechanical components. Medical Research Reactor (MRR). BNL 600 (T173), Brookhaven National Laboratory; Feb. 1960. Liu. H. B. Isodose curves and treatment planning for boron neutron capture therapy. Dissertation in Nuclear Engineering Program of University of Missouri, Columbia: Aug. 1992. Saraf. S. K.; Kalef-Ezra. J.: Fairchild, R. G.: Laster, B. H.; Fiarman. S.: Ramsay. E. Epithermal beam development at the BMRR: Dosimetric evaluation. In: Harling, 0. K., Bernard, J. A.. Zamenhof, R., eds. Neutron beam design, development. and performance for neutron capture therapy. New York: Plenum Press: I990:307-3 16. Slatkin. D. N. A history of boron neutron capture therapy of brain tumors-Postulation of a brain radiation dose tolerance limit. Brain Il4:1609-1629: 1991. Watkins, P.: Constantine, G.: Stecher-Rasmussen, F.: Freudenreich. W.: Moss. R. L.: Ricchena. R. MCNP calculations for the design and characterization ofthe Petten BNCT epithermal neutron beam. In: Allen, B. J., Moore. D. E.. Harrington. B. V., eds. Progress in neutron capture therapy for cancer. New York: Plenum Press: 1992:71-73. Zamenhof. R.; Brenner. J.; Yanch. J.: Wazer, D.: MadocJones. H.: Saris, S.; Harling, 0. Treatment planning for neutron capture therapy of glioblastoma multiforme using an epithermal neutron beam from the MITR-II research reactor and Monte Carlo simulation. In: Allen B. J., Moore. D. E.. Harrington. B. V.. eds. Progress in neutron capture therapy for cancer. New York: Plenum Press: 1992: l73177.