Neutron beam experiments using nuclear research reactors: honoring the retirement of professor Bernard W. Wehring -I. 3. A Comparison of Neutron Beams for BNCT

This paper evaluates the potential of the Ohio State University (OSU) Research Reactor (OSURR) with a fission convertor plate (FCP) for clinical boron neutron capture therapy (BNCT). The evaluation uses design methods that were developed for the analysis of the OSU design of an accelerator-based neutron source (ABNS) for BNCT (hereafter called the OSU-ABNS); namely, the in-phantom neutron field assessment parameters, the treatment time (T) and the high-LET absorbed-dose to the tumor (DTumor), were calculated using MCNP. The paper compares an FCP epithermal neutron beam, which is based on the OSURR (hereafter called the OSURR-FCP) with the OSU-ABNS. For completeness, the comparison includes an alternative ABNS design, which was taken from the literature (hereafter called the 7LiF-Al2O3 ABNS), and the Brookhaven Medical Research Reactor (BMRR) epithermal neutron beam for BNCT (hereafter called the BMRR-ENB). The OSURR-FCP design consists of the OSURR, a fission plate, and a moderator/filter assembly. These components were modeled in MCNP. The OSURR is a 500-kW pool-type light water-cooled and moderated reactor that is reflected on two sides with graphite and uses a U3Si2-Al dispersion fuel. The fission plate and moderator/filter assembly, which were modeled, were identical to those specified by Liu. The goal of our analysis was not to perfect an FCP and moderator/filter assembly for the OSURR-FCP. Rather, the intent of our analysis was to determine if, using the FCP and moderator/filter assembly designed by Liu, the OSURR, operating at 100% power, could produce a beam of sufficient intensity to treat human patients with BNCT in a reasonable treatment time. T is the total time required for a BNCT treatment, including all treatment fractions. Since the total dose delivered to the tumor is limited by the tolerance of the surrounding normal tissue, T is defined as the time required to escalate the normal tissue RBE-dose to the tolerance of the normal brain. DTumor is the product of the high-LET absorbed-dose rate to the tumor and the treatment time. A difference between our design methods and the design methods of others should be mentioned. The difference is that we chose to not take credit for the absorbed dose from gamma rays contributing to the control of gliomas. This assumption does not mean that we have completely ignored gamma rays in the development of our neutron field assessment parameters T and DTumor. For example, the radiation damage to normal tissue from gamma rays is included in the calculation of the treatment time T. The values of T and DTumor depend upon the 10B concentration in blood, the product of the RBE and the compound factor for the boron absorbed dose, the number of treatment fractions, and the ratio of the 10B concentration in tumor to the 10B concentration in blood. The values, which were assumed for these parameters, are appropriate for BNCT with BSH as the capture agent. For this paper, it is less important to know the exact values of the parameters that were used than it is to appreciate that the values of the parameters that were assumed were identical for the analysis of each neutron source that was evaluated. We calculated T and DTumor in a 14 x 14 x 14 cm Lucite cube phantom. Their calculation required the calculation of several intermediate quantities, such as the neutron absorbed dose rate (D·n), the gamma-ray absorbed dose rate (D·γ), and the 10B specific absorbed dose rate (d·B) as a function of depth in the phantom. Values for D·n, D·γ, d·B were calculated for all the neutron sources other than the BMRR-ENB, using the MCNP Monte Carlo radiation transport code. For the BMRR-ENB, the values of D·n, D·γ, d·B were supplied to us by our colleagues at Brookhaven National Laboratory. MCNP was run in n/p (neutron/photon) mode. The neutron and photon fluxes were calculated in the phantom along the phantom centerline. Trac k length estimators were used to tally neutron and photon fluxes in small spheres (0.2-cm radius) located ∼0.1 cm apart along the phantom centerline. To reduce the computational time required to obtain flux values with acceptable errors (fractional uncertainty <10%), splitting was implemented as a variance-reduction technique in the MCNP calculation. The values of D·n, D·γ, and d·B as a function of depth in the phantom were fit with curves using the Kaleidagraph software package on the Macintosh computer. The resulting curve fits were used to calculate T and DTumor. The point of view that we have adopted in designing an ABNS for BNCT is that glioblastoma multiform is a whole-brain disease and that recurrences are most likely to arise in the tumor margins that receive the least dose. For a bilateral irradiation, this corresponds to the midline of the brain, and hence, in designing an ABNS for BNCT, we have considered the value of DTumor at 7 cm to be most important since the phantom that we used in our analysis has an axial thickness of 14 cm. The value of DTumor, at a depth of 7 cm, is presented for each neutron source along with the treatment time T. A major goal of our analysis was to determine if a 500-kW reactor with an FCP can produce a neutron field with sufficient intensity to allow a patient to be treated in an acceptable treatment time with adequate beam quality. According to our analysis, the answer to this question is 'yes', provided that the patient is treated with at least four fractions. Although the quality of the neutron field for the OSURR-FCP is slightly inferior to the quality of the neutron field for the OSU-ABNS, the neutron field quality of the OSURR-FCP is judged to be acceptable, on the basis of comparison with DTumor at a depth of 7 cm for the BMRR-ENB. (authors)