Feasibility of Prompt Gamma-Ray Imaging in BNCT: SPECT-Based MCNP Simulations of Boron and Hydrogen Capture Events

Milad Mohammedali; Wonho Lee

doi:10.14407/jrpr.2025.00311

Abstract

Background

Boron neutron capture therapy (BNCT) is a targeted radiotherapy that selectively destroys cancer cells by delivering compounds enriched in boron-10 (¹⁰B) that preferentially accumulate in tumors. When exposed to low-energy neutrons, ¹⁰B undergoes a nuclear reaction that produces high-linear energy transfer particles that kill adjacent tumor cells while sparing healthy tissues. Real-time monitoring of boron distribution and the neutron irradiation is critical to optimize treatment efficacy and minimize collateral damage.

Materials and Methods

We evaluated a single-photon emission computed tomography (SPECT) configuration for simultaneous monitoring of boron spatial distribution and neutron irradiation patterns. Monte Carlo simulations were performed with the Monte Carlo N-Particle version 6.2 (MCNP6.2) using a male Oak Ridge National Laboratory stylized phantom. The simulations modeled prompt gamma emissions at 0.478 MeV (from ¹⁰B neutron capture) and 2.2 MeV (from hydrogen neutron capture) and the detection response of a four-head cadmium zinc telluride (CZT) SPECT system.

Results and Discussion

The SPECT system localized regions of boron uptake in the phantom using 0.478 MeV emissions with good spatial localization. Although the system detected 2.2 MeV gamma rays from hydrogen neutron capture, reconstructed images from this energy showed diffuse distributions and did not provide clear directional information about the neutron source. These findings indicate strong performance for boron mapping but limited capability for precise neutron-source directionality using 2.2 MeV gamma emissions in the present configuration.

Conclusion

SPECT-based imaging with a CZT detector shows feasibility for real-time boron mapping in BNCT. However, further work on detector geometry, collimation, and reconstruction algorithms is needed to improve sensitivity at clinically relevant boron concentrations and to enhance directional sensitivity for neutron-field assessment.

Keywords: Boron Neutron Capture Therapy, Single-Photon Emission Computed Tomography, Monte Carlo Simulation, Boron Uptake Region, Prompt Gamma Ray, Neutron Irradiation Mapping

Introduction

Neutron capture therapy was first proposed in 1936 by Gordon Locher [1], who suggested exploiting the high neutron capture cross-section of boron-10 (¹⁰B) to destroy tumors through the emission of high-energy alpha particles. Early clinical trials for glioblastoma multiforme in the United States, led by Sweet [2], used sodium borate and the Brookhaven Graphite Research Reactor. Inconsistent boron uptake and severe side effects, including necrosis and refractory shock, led to the suspension of these initial programs in 1961 [2].

Boron neutron capture therapy (BNCT) has advanced rapidly since the 1990s. The use of epithermal neutrons improved tissue penetration and broadened clinical applications [3]. Accelerator-based neutron sources emerged in Japan in 2012, enabling hospital implementation, and regulatory approval for BNCT in head and neck cancers followed in 2020 [4]. Compact accelerators such as Hyperion (Hyperion Materials & Technologies) have driven the establishment of hospital-based BNCT centers; by 2025, over 10 centers were reported operational or under development [5].

BNCT proceeds in two steps. First, a ¹⁰B compound (commonly boronophenylalanine or sodium borocaptate) is administered and preferentially accumulates in tumor cells. Second, the target is irradiated with epithermal neutrons that thermalize in tissue. Thermal neutrons captured by ¹⁰B nuclei produce high-energy alpha particles and lithium-7 nuclei that deposit energy over subcellular ranges. The boron capture reaction also emits a 0.478 MeV prompt gamma photon. Neutron capture on hydrogen yields a 2.2 MeV photon. Signatures can inform treatment planning and dosimetry [6].

Despite BNCT’s potential, challenges remain in monitoring boron concentration and neutron flux during treatment. Accurate, real-time monitoring is essential to ensure therapeutic dose delivery to tumors while minimizing normal tissue exposure. Existing methods, such as positron emission tomography and prompt gamma radiation analysis, face limitations from background signals and restrictions on real-time information [7, 8].

Single-photon emission computed tomography (SPECT) has been investigated to image the 0.478 MeV gamma from ¹⁰B capture. Still, detection of these gamma rays is complicated by interference from competing signals (for example, 0.511 MeV annihilation photons) and by low boron concentrations in tissues [9, 10].

Hydrogen capture (2.2 MeV) emissions are emitted approximately isotropically and can provide complementary information on the neutron field. Combining analysis of both 0.478 MeV and 2.2 MeV prompt gammas may improve real-time dosimetry and treatment accuracy. This study evaluates a cadmium zinc telluride (CZT)-based four-head SPECT system in Monte Carlo N-Particle version 6.2 (MCNP6.2; Los Alamos National Laboratory) simulations to assess feasibility for boron mapping and for neutron-field assessment in BNCT.

Materials and Methods

1. Source Configuration

An epithermal neutron source was modeled to provide neutron energies suitable for deep tissue penetration and effective tumor targeting. The neutron source was positioned 30 cm from the phantom center and emitted neutrons with energies between 0.5 eV and 10 keV. Those neutrons thermalize in tissue where they may be captured by ¹⁰B and hydrogen neutrons, producing prompt gamma emissions. The neutron spectrum was chosen to match previously reported BNCT source parameters [11]. Fig. 1 shows the neutron fluence spectrum.

2. System Design

Monte Carlo simulations were performed with MCNP6.2 to model neutron–photon transport and prompt gamma emission. The imaging system was a four-head SPECT geometry, as shown in Fig. 2. Each head used a high-resolution CZT detector with an active area of 20 cm×20 cm, segmented into 1 cm×1 cm pixels, and a thickness of 2 cm. Detector material density was 5.78 g/cm³. Detector centers were located 20 cm from the phantom center at angular positions of 0°, 90°, 180°, and 270°. Each detector employed a parallel-hole lead collimator (density: 11.34 g/cm³) with a thickness of 0.5 cm and a height of 5 cm; collimator geometry was matched to detector pixels to optimize spatial and energy resolution for BNCT-specific gamma energies [12, 13].

CZT was selected for its high-energy resolution (<2% full width at half maximum at 0.478 MeV) and high detection efficiency at 0.478 MeV and 2.2 MeV [14]. Transport calculations used the coupled neutron–photon mode. The ENDF/B-VII.1 nuclear data libraries were used. Standard MCNP energy cutoffs were used. Each simulation followed 1×10⁹ particle histories to ensure adequate statistical precision for photon and neutron spectra.

Detector response was recorded using the F8 pulse height tally to obtain energy spectra and generate two-dimensional (2D) projection images. Energy windows were set to isolate the two key peaks: 0.478 MeV (0.466–0.489 MeV) and 2.2 MeV (2.145–2.255 MeV). Gaussian energy broadening was applied to model detector energy resolution. Projection data were acquired in 5° increments through a full 360° rotation, combining stationary multiangle detection with rotational sampling to enable precise reconstruction of the boron distribution within the brain.

3. Phantom Modeling

The Oak Ridge National Laboratory (ORNL) phantom, a stylized computational model based on constructive solid geometry, represented human anatomy for radiation dosimetry and imaging simulations [15, 16]. The boron uptake region (BUR) was modeled as a 1 cm diameter sphere within the brain to represent a tumor target. The tumor center was positioned approximately 1.55 cm from the brain center along the z-axis, at coordinates (0, 0, and 93 cm), given a brain center at (0, 0, and 91.45 cm). Tissue compositions and anatomical segmentation were applied as provided by the phantom to simulate dose distribution and imaging response.

Results and Discussion

The simulated gamma-ray energy spectrum (F8 tally) showed a prominent peak at 0.478 MeV corresponding to ¹⁰B neutron capture, directly indicating the presence of boron in the tumor and a 2.2 MeV peak from the prompt gamma emission resulting from neutron capture by hydrogen (Fig. 3). These energy windows were selected to isolate and analyze the key gamma-ray signals relevant for BNCT. Additional spectral features included single and double escape peaks and annihilation peaks. The system’s minimum detectable boron concentration was estimated at 40,000 ppm in tumor volume and 10 ppm in normal tissue. These detection limits exceed typical clinical tumor boron concentrations (approximately 30 ppm) [16], indicating limited sensitivity for clinical levels of boron with the present setup. This constraint is due to the combination of realistic boron concentrations, detector geometry, collimation, and the inherent limitations of the simulation configuration. Simulations that assume pure boron or artificially high boron content predict much higher detectability, which emphasizes the sensitivity challenge for realistic tissue concentrations [11].

Projection data from each detector were formed into 2D projection images (85×85) across multiple angles (0°–360° in 5° increments). These projections were reconstructed into three-dimensional gamma emission distributions using the maximum likelihood expectation maximization (MLEM) algorithm. The MLEM update for voxel j at iteration k+1 is,

(1)

fj(k+1)=fj(k)∑i=1naijΣj′=1maij′fj′(k)gi,

where fj(k) is the estimated activity in voxel j at iteration k, g_i is the measured count in the projection bin i, and a_ij is the system matrix element giving the probability that an emission from voxel j will be detected in bin i. Equation (1) updates each voxel iteratively by comparing measured and estimated projections to maximize the likelihood of the observed data.

The MLEM reconstruction was performed for 0.478 MeV and 2.2 MeV energy windows. The 0.478 MeV reconstructions provided clear spatial localization of the BUR within the brain (Fig. 4). In contrast, 2.2 MeV reconstructions, while indicating hydrogen capture events, did not yield clear directional mapping of the incident neutron beam; images were predominantly diffuse and lacked strong anisotropic features (Fig. 5). The isotropic emission characteristic of 2.2 MeV gammas and the current detector/collimator geometry limit directional sensitivity.

To rigorously evaluate spatial resolution, simulations were repeated with two 1 cm BURs at variable separations (4, 3, and 2.5 cm), representing a range of clinically relevant scenarios for tumor localization within brain tissue. Reconstructions overlaid on cross-sections showed that BURs separated by 4 cm and 3 cm were resolved distinctly. In contrast, at 2.5 cm separation, the two regions merged into a single, indistinguishable region (Fig. 6). These results indicate that the current system resolves tumor-like uptake regions reliably at separations ≥3 cm, with a resolution between 2.5 cm and 3 cm under the simulated conditions.

The CZT detectors’ energy resolution (<2% FWHM at 0.478 MeV) enabled discrimination between 0.478 MeV gammas and nearby 0.511 MeV annihilation photons, contributing to accurate boron localization. However, detection of 2.2 MeV gammas showed limited directional sensitivity, primarily because hydrogen neutron capture gammas are emitted nearly isotropically and because the current collimation and detector geometry were not optimized for high-energy directional mapping. The novelty of this work is the combined use of 0.478 MeV and 2.2 MeV signals in a SPECT framework to assess the feasibility for simultaneous boron mapping and neutron-field assessment. Limitations of this study include the use of an ORNL stylized male phantom and a spherical 1 cm tumor model; patient-specific voxel phantoms would better represent realistic anatomy and heterogeneous tumor shapes in further studies.

Conclusion

Monte Carlo simulations with MCNP6.2 and a four-head CZT SPECT system demonstrate feasibility for imaging boron distribution in BNCT via 0.478 MeV prompt gamma detection. The system localized a 1 cm spherical BUR with good accuracy under the simulated conditions. Although 2.2 MeV hydrogen capture gammas were detectable, their reconstructed images were diffuse and provided limited information on neutron source directionality. Spatial resolution tests showed that two 1 cm BURs were distinguishable at separations of 4 cm and 3 cm, demonstrating a clinically viable spatial resolution for boron detection. However, at a separation of 2.5 cm, the regions merged visually, indicating a resolution threshold below which tumor differentiation becomes impractical.

Future work should investigate optimized detector geometry, collimator design, and reconstruction algorithms, and should use voxel-based, patient-specific phantoms to improve sensitivity at clinically relevant boron concentrations and to enhance directional sensitivity for neutron-field tracking.

Article Information

Funding

This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (MSIT) (RS-2024-00334576).

Conflict of Interest

No potential conflict of interest relevant to this article was reported.

Ethical Statement

This article does not contain any studies with human participants or animals performed by any of the authors.

Data Availability

The data that support the findings of this study are available from the corresponding author, Lee W, upon reasonable request.

Author Contribution

Conceptualization: Lee W. Methodology: all authors. Data curation: Mohammedali M. Formal analysis: Mohammedali M. Supervision: Lee W. Funding acquisition: Lee W. Project administration: Lee W. Investigation: Mohammedali M. Visualization: Mohammedali M. Software: Mohammedali M. Writing - original draft: Mohammedali M. Writing - review & editing: Lee W. Approval of final manuscript: all authors.

References

1. Locher G. Biological effects and therapeutic possibilities of neutrons. Am J Roentgenol Radium Ther. 1936;36(1):1-13.

2. Sweet WH. The uses of nuclear disintegration in the diagnosis and treatment of brain tumor. N Engl J Med. 1951;245(23):875-878.

3. Barth RF, Vicente MG, Harling OK, Kiger WS 3rd, Riley KJ, Binns PJ, et al. Current status of boron neutron capture therapy of high grade gliomas and recurrent head and neck cancer. Radiat Oncol. 2012;7:146.

4. Vyzhigina BB, Dolgushin BI, Kropotov MA. Efficacy of boron neutron capture therapy for head and neck cancer: systematic review. J Oncol Diagn Radiol Radiother. 2025;8(3):18-26.

5. Zheng D, Han G, Lemus ODM, Podgorsak A, Webster M, Li F, et al. Boron neutron capture therapy: a technology-driven Renaissance. Cancers (Basel). 2026;18(3):498.

6. Malouff TD, Seneviratne DS, Ebner DK, Stross WC, Waddle MR, Trifiletti DM, et al. Boron neutron capture therapy: a review of clinical applications. Front Oncol. 2021;11:601820.

7. Gong C, Tang X, Fatemi S. A Monte Carlo study of SPECT in boron neutron capture therapy for a heterogeneous human phantom. Int J Radiat Res. 2018;16(1):33.

8. Sakai M, Tamaki S, Murata I, Parajuli RK, Matsumura A, Kubo N, et al. Experimental study on Compton camera for boron neutron capture therapy applications. Sci Rep. 2023;13(1):22883.

9. Mousavi MS, Rahmani F. Feasibility study on the use of SPECT for boron dose imaging in TRR based BNCT. Radiat Phys Eng. 2022;3(1):1-12.

10. Ramos Lopez D, Pugliese GMI, Iaselli G, Amoroso N, Gong C, Pascali V, et al. Study of alternative imaging methods for in vivo boron neutron capture therapy. Cancers (Basel). 2023;15(14):3582.

11. Kim M, Hong BH, Cho I, Park C, Min SH, Hwang WT, et al. Design of a scintillator-based prompt gamma camera for boron neutron capture therapy: comparison of SrI2 and GAGG using Monte Carlo simulation. Nucl Eng Technol. 2021;53(2):626-636.

12. Van Audenhaege K, Van Holen R, Vandenberghe S, Vanhove C, Metzler SD, Moore SC. Review of SPECT collimator selection, optimization, and fabrication for clinical and preclinical imaging. Med Phys. 2015;42(8):4796-4813.

13. Kacperski K, Switlik D, Pietrzak J. High sensitivity collimators for optimising lesion detection in SPECT images. arXiv 2018 Jun 29 [Preprint]. https://doi.org/10.48550/arXiv.1806.11497

14. Schlesinger TE, Toney JE, Yoon H, Lee EY, Brunett BA, Franks L, et al. Cadmium zinc telluride and its use as a nuclear radiation detector material. Mater Sci Eng R Rep. 2001;32(4–5):103-189.

15. Han EY, Bolch WE, Eckerman KF. Revisions to the ORNL series of adult and pediatric computational phantoms for use with the MIRD schema. Health Phys. 2006;90(4):337-356.

16. Lee T, Lee H, Lee W. Monitoring the distribution of prompt gamma rays in boron neutron capture therapy using a multiple-scattering Compton camera: a Monte Carlo simulation study. Nucl Instrum Methods Phys Res A. 2015;798:135-139.

Fig. 1

Neutron-beam fluence.

Fig. 2

Diagram of single-photon emission computed tomography system geometry. CZT, cadmium zinc telluride.

Fig. 3

Energy spectrum using F8 Tally data (Monte Carlo N-Particle version 6.2).

Fig. 5

Maximum likelihood expectation maximization reconstructed image (5^th iteration) for hydrogen: 2.2 MeV in (A) XZ and (B) YZ planes.

Fig. 4

Maximum likelihood expectation maximization reconstructed image (5^th iteration) for ¹⁰B: 0.478 MeV in (A) XY, (B) XZ, and (C) YZ planes.

Fig. 6

Maximum likelihood expectation maximization reconstructions (5^th iteration) in (A) XY and (B) XZ planes for two boron uptake regions separated by 4.0, 3.0, and 2.5 cm (left to right).