A time- and space-saving Monte Carlo simulation method using post-collimation generative adversarial network for dose calculation of an O-ring gantry Linac.

PURPOSE: This study explores the feasibility of employing Generative Adversarial Networks (GANs) to model the RefleXion X1 Linac. The aim is to investigate the accuracy of dose simulation and assess the potential computational benefits. METHODS: The X1 Linac is a new radiotherapy machine with a binary multi-leaf collimation (MLC) system, facilitating innovative biology-guided radiotherapy. A total of 34 GAN generators, each representing a desired MLC aperture, were developed. Each generator was trained using a phase space file generated underneath the corresponding aperture, enabling the generation of particles and serving as a beam source for Monte Carlo simulation. Dose distributions in water were simulated for each aperture using both the GAN and phase space sources. The agreement between dose distributions was evaluated. The computational time reduction from bypassing the collimation simulation and storage space savings were estimated. RESULTS: The percentage depth dose at 10 cm, penumbra, and full-width half maximum of the GAN simulation agree with the phase space simulation, with differences of 0.4 % ± 0.2 %, 0.32 ± 0.66 mm, and 0.26 ± 0.44 mm, respectively. The gamma passing rate (1 %/1mm) for the planar dose exceeded 90 % for all apertures. The estimated time-saving for simulating an plan using 5766 beamlets was 530 CPU hours. The storage usage was reduced by a factor of 102. CONCLUSION: The utilization of the GAN in simulating the X1 Linac demonstrated remarkable accuracy and efficiency. The reductions in both computational time and storage requirements make this approach highly valuable for future dosimetry studies and beam modeling.

Small-field measurement and Monte Carlo model validation of a novel image-guided radiotherapy system.

PURPOSE: The RefleXion™ X1 is a novel radiotherapy system that is designed for image-guided radiotherapy, and eventually, biology-guided radiotherapy (BgRT). BgRT is a treatment paradigm that tracks tumor motion using real-time positron emission signals. This study reports the small-field measurement results and the validation of a Monte Carlo (MC) model of the first clinical RefleXion unit. METHODS: The RefleXion linear accelerator (linac) produces a 6 MV flattening filter free (FFF) photon beam and consists of a binary multileaf collimator (MLC) system with 64 leaves and two pairs of y-jaws. The maximum clinical field size achievable is 400 × 20 mm2 . The y-jaws provide either a 10 or 20 mm opening at source-to-axis distance (SAD) of 850 mm. The width of each MLC leaf at SAD is 6.25 mm. Percentage depth doses (PDDs) and relative beam profiles were acquired using an Edge diode detector in a water tank for field sizes from 12.5 × 10 to 100 × 20 mm2 . Beam profiles were also measured using films. Output factors of fields ranging from 6.25 × 10 to 100 × 20 mm2 were measured using W2 scintillator detector, Edge detector, and films. Output correction factors k of the Edge detector for RefleXion were calculated. An MC model of the linac including pre-MLC beam sources and detailed structures of MLC and lower y-jaws was validated against the measurements. Simulation codes BEAMnrc and GATE were utilized. RESULTS: The diode measured PDD at 10 cm depth (PDD10) increases from 53.6% to 56.9% as the field opens from 12.5 × 10 to 100 × 20 mm2 . The W2-measured output factor increases from 0.706 to 1 as the field opens from 6.25 × 10 to 100 × 20 mm2 (reference field size). The output factors acquired by diode and film differ from the W2 results by 1.65% (std = 1.49%) and 2.09% (std = 1.41%) on average, respectively. The profile penumbra and full-width half-maximum (FWHM) measured by diode agree well with the film results with a deviation of 0.60 mm and 0.73% on average, respectively. The averaged beam profile consistency calculated between the diode- and film-measured profiles among different depths is within 1.72%. By taking the W2 measurements as the ground truth, the output correction factors k for Edge detector ranging from 0.958 to 1 were reported. For the MC model validation, the simulated PDD10 agreed within 0.6% to the diode measurement. The MC-simulated output factor differed from the W2 results by 2.3% on average (std = 3.7%), while the MC simulated beam penumbra differed from the diode results by 0.67 mm on average (std = 0.42 mm). The MC FWHM agreed with the diode results to within 1.40% on average. The averaged beam profile consistency calculated between the diode and MC profiles among different depths is less than 1.29%. CONCLUSIONS: This study represents the first small-field dosimetry of a clinical RefleXion system. A complete and accurate MC model of the RefleXion linac has been validated.

Frequency-dependent optimal weighting approach for megavoltage multilayer imagers.

Multi-layer imaging (MLI) devices improve the detective quantum efficiency (DQE) while maintaining the spatial resolution of conventional mega-voltage (MV) x-ray detectors for applications in radiotherapy. To date, only MLIs with identical detector layers have been explored. However, it may be possible to instead use different scintillation materials in each layer to improve the final image quality. To this end, we developed and validated a method for optimally combining the individual images from each layer of MLI devices that are built with heterogeneous layers. Two configurations were modeled within the GATE Monte Carlo package by stacking different layers of a terbium doped gadolinium oxysulfide Gd2O2S:Tb (GOS) phosphor and a LKH-5 glass scintillator. Detector response was characterized in terms of the modulation transfer function (MTF), normalized noise power spectrum (NNPS) and DQE. Spatial frequency-dependent weighting factors were then analytically derived for each layer such that the total DQE of the summed combination image would be maximized across all spatial modes. The final image is obtained as the weighted sum of the sub-images from each layer. Optimal weighting factors that maximize the DQE were found to be the quotient of MTF and NNPS of each layer in the heterogeneous MLI detector. Results validated the improvement of the DQE across the entire frequency domain. For the LKH-5 slab configuration, DQE(0) increases between 2%-3% (absolute), while the corresponding improvement for the LKH-5 pixelated configuration was 7%. The performance of the weighting method was quantitatively evaluated with respect to spatial resolution, contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) of simulated planar images of phantoms at 2.5 and 6 MV. The line pair phantom acquisition exhibited a twofold increase in CNR and SNR, however MTF was degraded at spatial frequencies greater than 0.2 lp mm-1. For the Las Vegas phantom, the weighting improved the CNR by around 30% depending on the contrast region while the SNR values are higher by a factor of 2.5. These results indicate that the imaging performance of MLI systems can be enhanced using the proposed frequency-dependent weighting scheme. The CNR and SNR of the weighted combined image are improved across all spatial scales independent of the detector combination or photon beam energy.

GPU-accelerated Monte Carlo simulation of MV-CBCT.

Monte Carlo simulation (MCS) is one of the most accurate computation methods for dose calculation and image formation in radiation therapy. However, the high computational complexity and long execution time of MCS limits its broad use. In this paper, we present a novel strategy to accelerate MCS using a graphic processing unit (GPU), and we demonstrate the application in mega-voltage (MV) cone-beam computed tomography (CBCT) simulation. A new framework that generates a series of MV projections from a single simulation run is designed specifically for MV-CBCT acquisition. A Geant4-based GPU code for photon simulation is incorporated into the framework for the simulation of photon transport through a phantom volume. The FastEPID method, which accelerates the simulation of MV images, is modified and integrated into the framework. The proposed GPU-based simulation strategy was tested for its accuracy and efficiency in a Catphan 604 phantom and an anthropomorphic pelvis phantom with beam energies at 2.5 MV, 6 MV, and 6 MV FFF. In all cases, the proposed GPU-based simulation demonstrated great simulation accuracy and excellent agreement with measurement and CPU-based simulation in terms of reconstructed image qualities. The MV-CBCT simulation was accelerated by factors of roughly 900-2300 using an NVIDIA Tesla V100 GPU card against a 2.5 GHz AMD Opteron™ Processor 6380.

A rapid, accurate image simulation strategy for mega-voltage cone-beam computed tomography.

Intensive computation time is required to simulate images of electronic portal imaging device (EPID) using Monte Carlo (MC) technique, limiting the development of applications associated with EPID, such as mega-voltage cone-beam computed tomography (MV-CBCT). In this study, a fast, accurate simulation strategy for MV-CBCT utilizing the FastEPID technique has been developed and validated. During FastEPID simulation, photon detection was determined by pre-calculated photon energy deposition efficiency (η) and particle transport within the EPID was replaced with a pre-calculated optical photon spread function. This method is capable of reducing the time required for EPID image simulation by a factor of 90-140, without compromising image quality. MV-CBCT images reconstructed from the FastEPID simulated projections have been validated against measurement in terms of mean Hounsfield unit (HU), noise, and cupping artifact. These images were obtained with both a Catphan 604 phantom and an anthropomorphic pelvis phantom, under treatment beam energies of 2.5 MV, 6 MV, and 6 MV flattening filter free. The agreement between measurement and simulation was excellent in all cases. This novel strategy was capable of reducing the run time of a full scan simulation of MV-CBCT performed on a CPU cluster to a matter of hours, rather than weeks or months required by a conventional approach. Multiple applications associated with MV-CBCT (e.g. imager design optimization) are anticipated to gain from the implementation of this novel simulation strategy.

A novel method for fast image simulation of flat panel detectors.

We have developed a novel method for fast image simulation of flat panel detectors, based on the photon energy deposition efficiency and the optical spread function (OSF). The proposed method, FastEPID, determines the photon detection using photon energy deposition and replaces particle transport within the detector with precalculated OSFs. The FastEPID results are validated against experimental measurement and conventional Monte Carlo simulation in terms of modulation transfer function (MTF), signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), contrast, and relative difference of pixel value, obtained with a slanted slit image, Las Vegas phantom, and anthropomorphic pelvis phantom. Excellent agreement is observed between simulation and measurement in all cases. Without degrading image quality, the FastEPID method is capable of reducing simulation time up to a factor of 150. Multiple applications, such as imager design optimization for planar and volumetric imaging, are expected to benefit from the implementation of the FastEPID method.

A Monte Carlo study of the impact of phosphor optical properties on EPID imaging performance.

We have developed a Monte Carlo computational model of a clinically employed electronic portal imaging device (EPID), and demonstrated the impact of phosphor optical properties on the imaging performance. The EPID model was built with Geant4 application for tomographic emission. Both radiative and optical transport were included in the model. Modulation transfer function (MTF), normalized noise-power spectrum times the incident x-ray fluence (qNNPS), and detective quantum efficiency (DQE) were calculated for simulated and measured data, and their agreement was quantified by the normalized root-mean-square error (NRMSE). MTF was computed using a 100 µm wide slit tilted by 1.5° and qNNPS was estimated using the Fujita-Lubberts-Swank method. DQE was calculated from MTF and qNNPS data. The NRMSE value was 0.0467 for MTF, 0.0217 for qNNPS, and 0.0885 for DQE, showing good agreement between measurement and simulation. Five major optical properties, phosphor grain size, phosphor thickness, phosphor refractive index, binder refractive index, and packing ratio were tested for their influence on the qNNPS, MTF, and DQE(0) of the model. Generally, the effect on the qNNPS is greater than MTF, and no impact on DQE(0), except from phosphor thickness, was observed. Multiple applications, such as imager design optimization and investigations of the dosimetric performance, are expected to benefit from the validated model.