The sensitive element of the detector is a cylinder with diameter of 1 mm and length of 1 mm made of cast polystyrene scintillator (Bykov et al., 2021). It is placed inside of a light-reflecting cylinder made of hard Teflon with wall thickness of 1 mm. The top side of the scintillator is painted with custom-made MgO-based white paint. The scintillator is glued to a plastic optical fiber with epoxy, and all components are placed inside of a black plastic case. Two sensors, one with a boron enriched scintillator and a boron-free one, are combined in one detector head. Simultaneous application of two different registration channels: the first one, sensitive to γ-radiation, and the second one, sensitive to γ-radiation and neutrons, allows us to estimate the contribution of the neutron component accurately.

The idea to measure fast neutron dose and nitrogen dose is the following (Taskaeva and Taskaev, 2021). The cell lines are exposed to γ-radiation and mixed radiation (neutrons and γ-radiation) measuring the γ-ray dose D _γ. The γ-ray doses that cause the same effect, for example, cell survival, are compared. When cells are irradiated with γ-radiation, only a γ-ray dose is received, when they are irradiated with mixed radiation, a fast neutron dose and a nitrogen dose are added (there is no boron dose, since the boron drug is not used). The equivalent dose of high-LET particles (sum of the fast neutron dose and the nitrogen dose) is calculated by the formula: D _n = D _{γ standard}–D _{γ mixed}, where D _n–the equivalent dose of high-LET particles; D _{γ standard}–the dose of γ-radiation when the cells are exposed to γ-radiation; D _{γ mixed}–the dose of γ-radiation when the cells are exposed to mixed radiation.

It is well known (Wang et al., 2010; Faião-Flores et al., 2013; Yura and Fujita, 2013), that cell survival depends on the cell line culture, the absorbed dose, the type of ionizing radiation and its energy spectrum, time of the irradiation. The proposed method of dose measurement is free from the need to take into account all these factors. In the method, cell cultures placed in the same place are irradiated with two different types of ionizing radiation (γ-radiation and mixed radiation) for the same time and achieve the same survival rate. If irradiation of cell cultures of the same line with two different types of radiation for the same time leads to the same cell survival, then the equivalent doses are equal.

The principle of prompt γ-ray spectroscopy is based on γ-ray spectroscopy following neutron capture in ¹⁰B. In 93.9% of cases, a¹⁰B(n,α)⁷Li reaction proceeds with the emission of a recoiling ⁷Li nuclei in an excited state. Excited nuclei go to the ground state by the emission of 478 keV photons. The emission rate of the photons is proportional to the neutron capture reaction rate. This method is a direct method for measuring boron dose (Kobayashi and Kanda, 1983).

The complexity of the implementation is that the γ-ray spectrometer must be in the neutron flux, have a good energy resolution, while photons with the same energy are emitted from the lithium target as a result of inelastic scattering of protons on lithium atomic nuclei. If we use a γ-spectrometer, that is, relatively stable in the neutron flux, then the energy resolution does not allow us to separate the 478 keV line from the more powerful 511 keV line. The HPGe γ-ray spectrometer separates these lines, but it is not resistant to neutrons.

The principle of the epithermal neutron flux monitor is based on the activation method using ⁷¹Ga(n,γ)⁷²Ga reaction (Guan et al., 2015). In the monitor, the activation material is positioned in the geometrical center of the polymethyl methacrylate (PMMA) cylinder used for neutron moderation and covered with Cd foil for thermal neutron absorption. Gallium foil has a diameter of about 10 mm and a mass of about 50 mg; PMMA cylinder is 63 mm in height and 65 mm in diameter; Cd foil thickness is 0.1 mm. The detector is made collapsible in order to remove activated gallium foil and measure its activation by the HPGe γ-ray spectrometer. Of course, it is possible not to disassemble the detector. However, in this case, an intense cadmium emission line gives an additional load to the detector, which leads to an increase in the dead time of the spectrometer and to a systematic error in the intensity of γ-radiation measurement.

This monitor design allows you to achieve a flat sensitivity curve in the epithermal neutron energy range, while its sensitivities to fast and thermal neutrons are low.

Neutron irradiation was performed at the accelerator based neutron source VITA in the Budker Institute of Nuclear Physics, Novosibirsk, Russia (Taskaev et al., 2021a). The layout of the facility is shown in Figure 1.

The neutron source comprises a tandem electrostatic particle accelerator of original design (named Vacuum Insulated Tandem Accelerator–VITA) to produce a beam of protons or deuterons, thin solid lithium target to generate neutrons, a neutron beam shaping assembly to form a neutron beam. It is placed in two bunkers. The facility has the ability to place a lithium target in 5 positions; in Figure 1 they are marked as positions A, B, C, D, E. In these experiments, the lithium target was placed in positions A, C, or D.

In the accelerator, the energy of protons can be changed within a range of 0.6 MeV–2.3 MeV keeping a high-energy stability 0.1%. The beam current can also be changed in a range of 0.3–10 mA keeping a high current stability 0.4%.

Lithium target 10 cm in diameter has three layers: a thin layer of pure lithium to generate neutrons in ⁷Li(p,n)⁷Be reactions; a thin layer of material totally resistant to radiation blistering; and a thin copper substrate for efficient heat removal. This target provides a stable neutron yield for a long time with an acceptably low level of contamination of the beam transport path by the inevitably formed radioactive isotope beryllium-7.

A Beam Shaping Assembly (BSA) is applied to convert neutron flux into a beam of epithermal neutrons with characteristics suitable for clinical applications (Zaidi et al., 2018). The BSA consists of a magnesium fluoride moderator 210 mm thick, a composite reflector (graphite in the front hemisphere and lead in the back), an absorber, and a filter.

A polymethyl methacrylate (PMMA) moderator 72 mm thick is used to produce a thermal neutron beam suitable for irradiation of cell cultures and laboratory animals for development the BNCT technique. Compared to MgF₂ BSA this moderator provides a higher neutron flux, but slightly worse therapeutic ratio and depth of therapy.

The spatial distribution of the boron dose and the γ-ray dose in air and in a water phantom was measured by the small-sized boron neutron detector with a pair of cast polystyrene scintillators, one of which is enriched with boron.

For the treatment of patients, a BSA with a magnesium fluoride moderator and a composite reflector (graphite in the front hemisphere, lead in the back) was developed and manufactured. The results of simulating and measuring the dependence of the dose component rate considered in BNCT on the depth in the phantom using this BSA are shown in Figure 2. The energy of the proton beam is 2.2 MeV, the current is 1 mA. The phantom is a cube with a side of 15 cm, placed with 5 cm clearance from the moderator. In simulations, the phantom is filled with tissue-equivalent contents; in measurements, the phantom is filled with water. Measurements of boron dose and γ-ray dose were carried out by the developed boron neutron detector. The boron concentration is assumed to be 40 ppm. Figure 2 also shows the depth distribution of the so-called therapeutic coefficient TC–the ratio of the “useful” dose (boron dose) to the “harmful” dose (the sum of the γ-ray dose, the fast neutron dose and the nitrogen dose).

It can be seen that the formed beam of epithermal neutrons penetrates deep enough, providing the maximum boron dose at a depth of 2 cm and 1/e times less at a depth of 6.6 cm. If the phantom is placed close to the moderator, then the boron dose increases by 1.5 times, but the therapeutic coefficient is reduced by 8%. Since simulations show that an increase in the proton energy from 2.2 to 2.3 MeV leads to an increase in both the boron dose and the therapeutic coefficient, it is planned to test this experimentally after replacing the accelerator electrodes damaged by operation at the extremely high parameters with new ones.

The facility also uses a PMMA moderator, which ensures an acceptable neutron beam quality when irradiating cell cultures and laboratory animals, when testing targeted boron delivery drugs, and when treating domestic animals with spontaneous tumors.

The results of simulating and measuring the dependence of the dose component rate on depth in a phantom placed close to a 72 mm thick PMMA moderator are shown in Figure 3. In this case, the energy of the proton beam is 2.1 MeV, the current is 1 mA. Since the neutron spectrum is shifted to the region of thermal energies due to the hydrogen-containing moderator, the maximum boron dose is realized on the surface and the neutrons penetrate into the phantom by a smaller one–e times the boron dose rate drops at a depth of 3.4 cm. It is also seen that the therapeutic coefficient of such the neutron beam is 1.5 times smaller. However, such a beam provides 2.7 times the dose rate at a proton energy of 2.1 MeV and 4.4 times at an energy of 2.2 MeV compared to a magnesium fluoride moderator, if the phantom is also placed close to the moderator.

To improve the quality of the neutron beam using a PMMA moderator, it is proposed to make the moderator thicker in the forward direction–this will reduce the fast neutron dose, and to sprinkle bismuth inside the PMMA–this will reduce the contribution of the γ-ray dose. Such a change, as shown by the simulation results, will improve the therapeutic coefficient by a third–up to 6, with a slight decrease in the boron dose rate.

The glioblastoma cell line U251 and the clonogenic test were used to implement the dose measurement method. At first, U251 cells were irradiated only with γ-radiation and their survival was determined. A γ-radiation was produced in the ⁷Li(p,p’γ)⁷Li reaction at a proton beam energy below 1.882 MeV–threshold of the ⁷Li(p,n)⁷Be reaction. The photon energy in this reaction is 478 keV. U251 cell line was irradiated for 1.5 h with photons at proton energy 1.800 ± 0.002 MeV and proton current 2.17 ± 0.03 mA. The dose absorbed by the cells was 5.21 Gy. After γ-radiation the surviving fractions of U251 cell line were examined using a clonogenic assay. Cell survival as a result of photon irradiation was 34% ± 4%.

To generate mixed radiation, neutron and γ-radiation, the proton energy was increased to 2.05 MeV–above the threshold of the ⁷Li(p,n)⁷Be reaction, the proton beam current was reduced. The experiment was carried out according to the following scenario. A certain value of the proton beam current was set, the cells were irradiated for 1.5 h, and then their survival was determined. If the cell survival is more than 34%, then in the next irradiation the beam current will be increased, if it is less it will be reduced. After several attempts the cell survival fraction was achieved exactly 34%. The dose measured with a γ-dosimeter was 3.98 Gy.

Since the survival rate of cells after irradiation with mixed radiation is the same as that of γ-radiation and the irradiation times were equal, the cells received the same equivalent dose—5.21 Gy-Eq. When irradiated with mixed radiation, γ-radiation gave 3.98 Gy-Eq at this dose, which means that the remaining 1.23 Gy-Eq were due to recoil nuclei caused by the absorption of thermal neutrons and elastic scattering of fast neutrons.

In this experiment, the lithium target was placed in position D, and the HPGe γ-spectrometer in bunker 2 (see Figure 1). A 50 mL test tube with boron of various concentrations is placed in a neutron beam, and the γ-ray spectrum is measured with the HPGe γ-spectrometer within 15 min. The result obtained is shown in Figure 4. It can be seen that the 478 keV photon line broadened by the Doppler effect is reliably measured and the boron content is determined starting from a 20 ppm content. Measuring this concentration of boron is sufficient to estimate the boron dose during therapy. Note that the 511 keV photon line due to neutron absorption by hydrogen is separated from the 478 keV photon line and can be used to monitor the neutron flux.

The epithermal neutron flux monitor is placed along the proton beam axis at a distance of 100 mm from the lithium target placed in position A (see Figure 1). Discs made of PMMA with a diameter of 200 mm and a thickness of 12 mm are placed closely below the target in an amount of 0–6 pcs. The target is irradiated to a fluence of 2.5 C for 1 h at a proton energy of 2 MeV. After irradiation, the monitor is disassembled, the gallium foil is removed and the HPGe γ-spectrometer measures the activation of the gallium foil. The measured activation of the gallium foil, taking into account its mass, makes it possible to reconstruct the epithermal neutron flux density.

The results are shown in Figure 5. It can be seen that the experimental results are very different from the simulation results in the absence of disks and with a single disk 12 mm thick, but are consistent with for a PMMA thickness of more than 24 mm. The simulated neutron energy spectrum for several PMMA thicknesses is shown in Figure 6. It can be seen that without the use of a moderator, fast neutrons predominate in the energy spectrum of neutrons. Although the sensitivity of the monitor to fast neutrons is low, they make a significant contribution to the activation of gallium due to their large number. As the moderator thickness increases the energy spectrum shifts towards the epithermal energy range and the measurement results become closer to the simulation results.