Full Monte Carlo Simulation Research Articles

Novel signatures of radiation reaction in electron–laser sidescattering

In this article we investigate novel signatures of radiation reaction via the angular deflection of an electron beam colliding at 90 degrees with an intense laser pulse. Due to the radiation reaction effect, the electrons can be deflected towards the beam axis for plane wave backgrounds, which is not possible in the absence of radiation reaction effects. The magnitude and size of the deflection angle can be controlled by tailoring the laser pulse shapes. The effect is first derived analytically using the Landau–Lifshitz equation, which allows to determine the important scaling behavior with laser intensity and particle energy. We then move on to full scale 3D Monte Carlo simulations to verify the effect is observable with present day laser technology. We investigate the opportunities for an indirect observation of laser depletion in such side scattering scenarios.

Determination of kilovoltage x-ray therapy depth doses with open-ended applicators.

The purpose of this work was to determine percentage depth dose (PDD) curves for kilovoltage x-rays from the WOmed-T105 unit, with open-ended steel applicators and beam qualities ranging from 0.5 to 4.2mm Al. Measurements were made with parallel plate chambers in a water phantom, with extrapolation based on a fifth order polynomial used to estimate the surface dose. Measurements were also made with parallel plate chambers in a plastic water phantom, with thin plastic sheets used to obtain detailed measurements at shallow depths (less than 1mm). Monte Carlo simulations were performed using the EGSnrc package, with two different sources as input: a SpekPy simulation of the x-ray beam and a full simulation of the x-ray tube, treatment head and applicators. Results showed that all four methods (two measurements and two simulations) agreed within the measurement uncertainty at depths greater than 2mm. At shallow depths, significant differences were noted. At depths less than 0.1mm, the full Monte Carlo simulation and the solid water measurements showed a sharp spike in surface dose which is attributed to electron contamination, which was not seen in the SpekPy Monte Carlo simulation or the extrapolated water measurements. At depths between 0.1mm and 2mm, beyond the range of contaminant electrons, the extrapolated water measurements underestimate the dose by up to 13% compared to the full Monte Carlo simulation and the solid water measurements, attributed to fluorescent photons generated in the applicators. This work demonstrates that for open-ended applicators, measurement of depth doses in water with extrapolation of surface dose has the potential to significantly underestimate the dose at shallow depths between the surface and 2mm, even after eliminating electron contamination from the beam.

The H 0 trouble: confronting non-thermal dark matter and phantom cosmology with the CMB, BAO, and Type Ia supernovae data

We have witnessed different values of the Hubble constant being found in the literature in the past years. Albeit, early measurements often result in an H 0 much smaller than those from late-time ones, producing a statistically significant discrepancy, and giving rise to the so-called Hubble tension. The trouble with the Hubble constant is often treated as a cosmological problem. However, the Hubble constant can be a laboratory to probe cosmology and particle physics models. In our work, we will investigate if the possibility of explaining the H 0 trouble using non-thermal dark matter production aided by phantom-like cosmology is consistent with the Cosmic Background Radiation (CMB) and Baryon Acoustic Oscillation (BAO) data. We performed a full Monte Carlo simulation using CMB and BAO datasets keeping the cosmological parameters Ω bh 2, Ω ch 2, 100θ, τopt , and w as priors and concluded that a non-thermal dark matter production aided by phantom-like cosmology yields at most H 0 = 70.5 km s-1 Mpc-1 which is consistent with some late-time measurements. However, if H 0 > 72 km s-1 Mpc-1 as many late-time observations indicate, an alternative solution to the Hubble trouble is needed. Lastly, we limited the fraction of relativistic dark matter at the matter-radiation equality to be at most 1%.

A novel experimental search channel for very light higgs bosons in the 2HDM type I

We present a reinterpretation study of existing results from the CMS Collaboration, specifically, searches for light Beyond the Standard Model (BSM) Higgs pairs produced in the chain decay pp→HSM→hh(aa)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$pp\\rightarrow H_\ extrm{SM}\\rightarrow hh(aa)$$\\end{document} into a variety of final states, in the context of the CP-conserving 2-Higgs Doublet Model (2HDM) Type-I. Through this, we test the Large Hadron Collider (LHC) sensitivity to a possible new signature, pp→HSM→ZA→ZZh\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$pp\\rightarrow H_\ extrm{SM}\\rightarrow ZA\\rightarrow ZZ h$$\\end{document}, with ZZ→jjμ+μ-\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ZZ\\rightarrow jj \\mu ^+\\mu ^-$$\\end{document} and h→bb¯\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$h\\rightarrow b\\bar{b}$$\\end{document}. We perform a systematic scan over the 2HDM Type-I parameter space, by taking into account all available theoretical and experimental constraints, in order to find a region with a potentially visible signal. We investigate the significance of it through a full Monte Carlo simulation down to the parametrised detector level. We show that such a signal is an alternative promising channel to standard four-body searches for light BSM Higgs bosons at the LHC already with an integrated luminosity of L=300fb-1\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${{\\mathcal {L}}} = 300~\ extrm{fb}^{-1}$$\\end{document}. For a tenfold increase of the latter, discovery should be possible over most of the allowed parameter space.

EFFICIENT WRONG-WAY RISK MODELING FOR FUNDING VALUATION ADJUSTMENTS

Wrong-Way Risk (WWR) is an important component in Funding Valuation Adjustment (FVA) modeling. Yet, the standard assumption is independence between market risks and the counterparty defaults and funding costs. This typical industrial setting is our point of departure, where we aim to assess the impact of WWR without running a full Monte Carlo simulation with all credit and funding processes. We propose to split the exposure profile into two parts: an independent and a WWR-driven part. For the former, exposures can be re-used from the standard xVA calculation. We express the second part of the exposure profile in terms of the stochastic drivers and approximate these by a common Gaussian stochastic factor. Within the affine setting, the proposed approximation is generic, is an add-on to the existing xVA calculations and provides an efficient and robust way to include WWR in FVA modeling. Case studies for an interest rate swap and a representative multi-currency portfolio of swaps illustrate that the approximation method is applicable in a practical setting. We analyze the approximation error and use the approximation to compute WWR sensitivities, which are needed for risk management. The approach is equally applicable to other metrics such as Credit Valuation Adjustment.

High dose rate 192Ir brachytherapy source model Monte Carlo dosimetry: mHDR-v2 and mHDR-v2r.

After 2010, the source model of the microSelectron HDR Afterloader System was slightly modified from the previous model. Granero et al. named the modified source model "mHDR-v2r (revised model mHDR-v2)" and the previous model "mHDR-v2". They concluded that the dosimetric differences arising from the dimensional changes between the mHDR-v2 and mHDR-v2r designs were negligible at almost all locations (within 0.5% for r ≥ 0.25 cm), the two-dimensional anisotropy function difference between the two sources is found 2.1% at r = 1.0 cm when compared with the results of the other experimental group. To confirm this difference, we performed a full Monte Carlo simulation without energy-fluence approximation. This is useful near the radiation source where charged-particle equilibrium does not hold. The two-dimensional anisotropy function of the TG-43U1 dataset showed a few percent difference between the mHDR-v2r and mHDR-v2 sources. There was no agreement in the immediate vicinity of the source (0.10 cm and 0.25 cm), when compared to Granero et al. in mHDR-v2r sources. The differences in these two-dimensional anisotropy functions were identified.

Radioactivation effects of titanium caused by clinical proton beam: a simulation study

Objective. In proton beam therapy (PBT), metals in the patient body perturb the dose distribution, and their radioactivation may affect the dose distribution around the metal; however, the radioactivation effect has been not clarified with PBT. In this study, we aimed to evaluate the radioactivation effect of metal depending on proton energies and secondary neutrons with a clinical proton beam using a Monte Carlo (MC) simulation. Approach. The radionuclides produced from a titanium alloy (Ti-6Al-4V) and their radioactivity were calculated using a 210-MeV passive scattering proton beam with a 60-mm Spread-out Bragg Peak, and the deposited doses caused by the radioactivation were computed using the MC simulation. The position of metal was changed according to the proton mean energy in water. To assess neutron effects on the radioactivation, we calculated the radioactivation in following three situations: (i) full MC simulation with neutrons, (ii) simulation without secondary neutrons generated from the beamline components, and (iii) simulation without any secondary neutrons. Main results. Immediately after the irradiation, the radionuclide with the largest activity was Sc-45 m (half-life of 318 ms) regardless of the proton energy and the presence of neutrons. Total radioactivity tended to increase according to the proton energy. The accumulated dose for 24 h caused by the metal activation showed an increasing trend with the proton energy, with a maximum increase rate of 0.045% to the prescribed dose. The accumulated dose at a distance of 10 mm from the metal was lower than 1/10 of that at a distance of 1 mm. Significance. The radioactivation effect of the titanium was comprehensively evaluated in the clinical passive scattering proton beam. We expect that radioactivation effects on the clinical dose distribution would be small. We consider that these results will help the clinical handling of high-Z metals in PBT.

Upgrading the neutronic version of the PSI cycle check-up methodology for full core PWR Monte Carlo simulations

Full core Monte Carlo (MC) simulations are of high interest and can be used to study the beginning of live conditions of reactors. However, current hardware limitations make it difficult to perform such studies for other cycles, as the whole irradiation history of the fuel assemblies (FAs) loaded must be considered on a time scale of several years using MC depletion capabilities. Therefore, at PSI, a cycle check-up methodology (CHUP) is currently being developed to address this challenge. By extracting operating conditions and isotopic compositions from validated reference deterministic core-follow models (using CASMO5/SIMULATE5), MC neutron transport models are generated for selected operating points. These models can then be used to estimate quantities beyond the capabilities of deterministic solvers. This article presents the verification performed on a hot zero power (HZP) operating condition of a Swiss PWR using the newly updated CHUP methodology. This update consists of extracting information generated automatically from reference core models, minimising as much human intervention as possible. The performance of the method was first evaluated against the reference deterministic lattice code. Then, the memory requirements of the full core simulation were studied under different approximations. Finally, verification was performed against the nodal reference scheme.

Physics-Guided Deep Scatter Estimation by Weak Supervision for Quantitative SPECT.

Accurate scatter estimation is important in quantitative SPECT for improving image contrast and accuracy. With a large number of photon histories, Monte-Carlo (MC) simulation can yield accurate scatter estimation, but is computationally expensive. Recent deep learning-based approaches can yield accurate scatter estimates quickly, yet full MC simulation is still required to generate scatter estimates as ground truth labels for all training data. Here we propose a physics-guided weakly supervised training framework for fast and accurate scatter estimation in quantitative SPECT by using a 100× shorter MC simulation as weak labels and enhancing them with deep neural networks. Our weakly supervised approach also allows quick fine-tuning of the trained network to any new test data for further improved performance with an additional short MC simulation (weak label) for patient-specific scatter modelling. Our method was trained with 18 XCAT phantoms with diverse anatomies / activities and then was evaluated on 6 XCAT phantoms, 4 realistic virtual patient phantoms, 1 torso phantom and 3 clinical scans from 2 patients for 177Lu SPECT with single / dual photopeaks (113, 208 keV). Our proposed weakly supervised method yielded comparable performance to the supervised counterpart in phantom experiments, but with significantly reduced computation in labeling. Our proposed method with patient-specific fine-tuning achieved more accurate scatter estimates than the supervised method in clinical scans. Our method with physics-guided weak supervision enables accurate deep scatter estimation in quantitative SPECT, while requiring much lower computation in labeling, enabling patient-specific fine-tuning capability in testing.

Beyond lepton number violation at the HL-LHC: resolving heavy neutrino-antineutrino oscillations

Collider testable low-scale seesaw models predict pseudo-Dirac heavy neutrinos, that can produce an oscillating pattern of lepton number conserving and lepton number violating events. We explore if such heavy neutrino-antineutrino oscillations can be resolved at the HL-LHC. To that end, we employ the first ever full Monte Carlo simulation of the oscillations, for several example benchmark points, and show under which conditions the CMS experiment is able to discover them. The workflow builds on a FeynRules model file for the phenomenological symmetry protected seesaw scenario (pSPSS) and a patched version of MadGraph, able to simulate heavy neutrino-antineutrino oscillations. We use the fast detector simulation Delphes and present a statistical analysis capable of inferring the significance of oscillations in the simulated data. Our results demonstrate that, for heavy neutrino mass splittings smaller than about 100 μeV, the discovery prospects for heavy neutrino-antineutrino oscillation at the HL-LHC are promising.

Monte Carlo investigation of dose distribution of uniformly and non-uniformly loaded standard and notched eye plaques.

To investigate the effect of using non-uniform loading and notched plaques on dose distribution for eye plaques. Using EGSnrc Monte Carlo (MC) simulations, we investigate eye plaque dose distributions in water and in an anatomically representative eye phantom. Simulations were performed in accordance with TG43 formalism and compared against full MC simulations which account for inter-seed and inhomogeneity effects. For standard plaque configurations, uniformly and non-uniformly loaded plaque dose distributions in water showed virtually no difference between each other. For standard plaque, the MC calculated dose distribution in planes parallel to the plaque is narrower than the TG43 calculation due to attenuation at the periphery of the plaque by the modulay. MC calculated the dose behind the plaque is fully attenuated. Similar results were found for the notched plaque, with asymmetric attenuation along the plane of the notch. Cumulative dose volume histograms showed significant reductions in the calculated MC doses for both tumor and eye structures, compared to TG43 calculations. The effect was most pronounced for the notch plaque where the MC dose to the optic nerve was greatly attenuated by the modulay surrounding the optic nerve compared to the TG43. Thus, a reduction of optic nerve D95% from 14 to 0.2Gy was observed, when comparing the TG43 calculation to the MC result. The tumor D95% reduced from 89.2 to 79.95Gy for TG43 and MC calculations, respectively. TG43 calculations overestimate the absolute dose and the lateral dose distribution of both standard and notched eye plaques, leading to the dose overestimation for the target and organs at risk. The dose matching along the central axis for the non-uniformly loaded plaques to that of uniformly loaded ones was found to be sufficient for providing comparable coverage and can be clinically used in eye-cancer-busy centers.

EASpy: fast simulation of fluorescence and Cherenkov light from extended air showers at large zenith angles

The detailed simulation of extended air showers (EAS) and their emission of Cherenkov and fluorescence light requires increasing computation time and storage volume with increasing energy of the primary particle. Given these limitations, it is currently challenging to optimize configurations of imaging air Cherenkov telescopes at photon energies beyond approximately 100 TeV. Additionally, the existing simulation frameworks are not capable of capturing the interplay of Cherenkov and fluorescence light emission at large zenith angle distances (≳ 70°), where the collection area of Cherenkov telescopes considerably increases. Here, we present EASpy, a framework for the simulation of EAS at large zenith angles using parametrizations for electron-positron distributions. Our proposed approach for the emission of fluorescence and Cherenkov light and the subsequent imaging of these components by Imaging Atmospheric Cherenkov Telescopes (IACTs) aims to provide flexibility and accuracy while at the same time it reduces the computation time considerably compared to full Monte Carlo simulations. We find excellent agreement of the resulting Cherenkov images when comparing results obtained from EASpy with the de-facto standard simulation tool CORSIKA and sim_telarray. In the process of verifying our approach, we have found that air shower images appear wider and longer with increasing impact distance at large zenith angles, an effect that has previously not been noted. We also investigate the distribution of light on the ground for fluorescence and Cherenkov emission and highlight their key differences to distributions at moderate zenith angles.

A numerical study of the spectral radiative properties of packed bed with mixed bauxite and silica spheres

Bauxite and silica particles have gained increasing attention for applications in the field of concentrated solar power. In this work, a Monte Carlo ray-tracing simulation is performed to predict the radiative properties (absorptance, reflectance, and transmittance) of packed beds with mixed bauxite and silica spherical particles at wavelengths of 0.5, 2.6, 9.0, and 9.35 µm. These wavelengths are representative for the visible, near-infrared, and the mid-infrared regions that are important for solar and thermal radiation, respectively. A repeating unit-column approach is used to mathematically represent the particle bed. The effects of particle mixing ratio, volume fraction, and wavelength on the predicted radiative properties are examined. The obtained radiative properties are inputted to an inverse method to retrieve the effective absorption and scattering coefficients as well as the scattering albedo, which may be later used in a continuous-scale radiative heat transfer analysis. Furthermore, the independent scattering model is used to obtain the absorption coefficient and scattering albedo based on the absorption and scattering cross sections predicted by a Monte Carlo algorithm for a single particle. It is shown that the independent scattering model underpredicts the scattering coefficient for opaque particles but overpredicts the scattering coefficient for semitransparent particles for sufficiently high particle volume fractions. The radiative properties calculated from the independent scattering model are compared to the full Monte Carlo simulation of the particle bed to examine the influence of particle mixing on dependent scattering.

A dose calculation algorithm for radiosurgery treatment room shielding optimization.

Methods described in the literature for designing shielding for treatment rooms for radiotherapy systems often involve assumptions that lead to overestimations of the wall thicknesses required to meet dose rate constraints outside the room. The Leksell Gamma Knife (LGK) has built-in shielding that results in primarily scattered photons leaking into the room. The field of leakage radiation, therefore, has a wide spectrum of energies, up to the primary energies of cobalt-60, and is highly anisotropic, making standard site planning methods difficult to adapt to the LGK. Monte Carlo (MC) simulations are an alternative for dose calculations but are computationallyexpensive. The aim was to develop a dose calculation algorithm for fast estimations of dose rates outside the barriers of a treatment room. The algorithm could then be employed in iterative methods to optimize treatment room parameters such as wall thicknesses and position of the LKGunit. The algorithm uses pre-calculated MC simulation data in two steps. First, it uses phase spaces that describe single photons in the radiation field around the LGK. Using each photon's position and direction, they are raytraced to the outside of the room. Based on their energy, angle of incidence to the barrier, and the barrier's thickness, a depth-dose profile is sampled from a pre-calculated library of profiles. The contribution from each photon is added to the space outside the barriers, giving the total dose distribution. The barrier thicknesses are optimized by iteratively running the algorithm and finding the minimum thickness required to keep the resulting maximum dose rate outside below a given upperlimit. Comparisons between dose distributions from the dose algorithm and full MC simulations of two typical rooms show good agreement, with the maximum dose rate outside each barrier being estimated within what represents a 2 cm error in concrete barrier thickness. The algorithm is successfully used in optimizing the barrier thicknesses and the results show potential decreases in most barriers' thickness, compared to a conventionalmethod. The algorithm is a promising alternative to conventional barrier design methods, eliminating several shortcomings of the latter whilst being significantly faster than a full MC simulation. The flexibility of the algorithm would further enable solving more complex cases, such as optimizing the LGK position or barrier material to further reduce material use andcost.

Ferromagnetic domains in the large- U Hubbard model with a few holes: A full configuration interaction quantum Monte Carlo study

Two-dimensional Hubbard lattices with two or three holes are investigated as a function of $U$ in the large-$U$ limit. In the so-called Nagaoka limit (one-hole system at infinite $U$), it is known that the Hubbard model exhibits a ferromagnetic ground state. Here, by means of exact full configuration interaction quantum Monte Carlo simulations applied to periodic lattices up to 24 sites, we compute spin-spin correlation functions as a function of increasing $U$. The correlation functions clearly demonstrate the onset of ferromagnetic domains, centered on individual holes. The overall total spin of the wave functions remains the lowest possible (0 or $\frac{1}{2}$, depending on the number of holes). The ferromagnetic domains appear at interaction strengths comparable to the critical interaction strengths of the Nagaoka transition in finite systems with strictly one hole. The existence of such ferromagnetic domains is the signature of Nagaoka physics in Hubbard systems with a small (but greater than 1) number of holes.

Terzina on board NUSES: A pathfinder for EAS Cherenkov Light Detection from space

In this paper we introduce the Terzina telescope as a part of the NUSES space mission. This telescope aims to detect Ultra High Energy Cosmic Rays (UHECRs) through the Cherenkov light emission from the extensive air showers (EAS) that they create in the Earth’s atmosphere. The Cherenkov photons are aligned along the shower axis inside about ∼ 0.2 − 1°, so that they become detectable by Terzina when it points towards the Earth’s limb. A sun-synchronous orbit will allow the telescope to observe only the night side of the Earth’s atmosphere. In this contribution, we focus on the description of the telescope detection goals, geometry, optical design and its photon detection camera composed of Silicon Photo-Multipliers (SiPMs). Moreover, we describe the full Monte Carlo simulation chain developed to estimate Terzina’s performance for UHECR detection. The estimate of the radiation damage and light background rates, the readout electronics and trigger logic are briefly described. Terzina will be able to study the potential for future physics missions devoted to UHECR detection and to UHE neutrino astronomy. It is a pathfinder for missions like POEMMA or future constellations of similar satellites to NUSES.

Prediction of DNA rejoining kinetics and cell survival after proton irradiation for V79 cells using Geant4-DNA

Purpose:Track structure Monte Carlo (MC) codes have achieved successful outcomes in the quantitative investigation of radiation-induced initial DNA damage. The aim of the present study is to extend a Geant4-DNA radiobiological application by incorporating a feature allowing for the prediction of DNA rejoining kinetics and corresponding cell surviving fraction along time after irradiation, for a Chinese hamster V79 cell line, which is one of the most popular and widely investigated cell lines in radiobiology. Methods:We implemented the Two-Lesion Kinetics (TLK) model, originally proposed by Stewart, which allows for simulations to calculate residual DNA damage and surviving fraction along time via the number of initial DNA damage and its complexity as inputs. Results:By optimizing the model parameters of the TLK model in accordance to the experimental data on V79, we were able to predict both DNA rejoining kinetics at low linear energy transfers (LET) and cell surviving fraction. Conclusion:This is the first study to demonstrate the implementation of both the cell surviving fraction and the DNA rejoining kinetics with the estimated initial DNA damage, in a realistic cell geometrical model simulated by full track structure MC simulations at DNA level and for various LET. These simulation and model make the link between mechanistic physical/chemical damage processes and these two specific biological endpoints.

Clinical treatment planning for kilovoltage radiotherapy using EGSnrc and Python.

Kilovoltage radiotherapy dose calculations are generally performed with manual point dose calculations based on water dosimetry. Tissue heterogeneities, irregular surfaces, and introduction of lead cutouts for treatment are either not taken into account or crudely approximated in manual calculations. Full Monte Carlo (MC) simulations can account for these limitations but require a validated treatment unit model, accurately segmented patient tissues and a treatment planning interface (TPI) to facilitate the simulation setup and result analysis. EGSnrc was used in this work to create a model of Xstrahl kilovoltage unit extending the range of energies, applicators, and validation parameters previously published. The novel functionality of the Python-based framework developed in this work allowed beam modification using custom lead cutouts and shields, commonly present in kilovoltage treatments, as well as absolute dose normalization using the output of the unit. 3D user-friendly planning interface of the developed framework facilitated non-co-planar beam setups for CT phantom MC simulations in DOSXYZnrc. The MC models of 49 clinical beams showed good agreement with measured and reference data, to within 2% for percentage depth dose curves, 4% for beam profiles at various depths, 2% for backscatter factors, 0.5 mm of absorber material for half-value layers, and 3% for outputfactors. End-to-end testing of the framework using custom lead cutouts resulted in good agreement to within 3% of absolute dose distribution between simulations and EBT3 GafChromic film measurements. Gamma analysis demonstrated poor agreement at the field edges which was attributed to the limitations of simulating smooth cutout shapes. Dose simulated in a heterogeneous phantom agreed to within 7% with measured values converted using the ratio of mass energy absorption coefficients of appropriate tissues andair.

Small beams, fast predictions: a comparison of machine learning dose prediction models for proton minibeam therapy.

Dose calculations for novel radiotherapy cancer treatments such as proton minibeam radiation therapy is often done using full Monte Carlo (MC) simulations. As MC simulations can be very time consuming for this kind of application, deep learning models have been considered to accelerate dose estimation in cancer patients. This work systematically evaluates the dose prediction accuracy, speed and generalization performance of three selected state-of-the-art deep learning models for dose prediction applied to the proton minibeam therapy. The strengths and weaknesses of those models are thoroughly investigated, helping other researchers to decide on a viable algorithm for their ownapplication. The following recently published models are compared: first, a 3D U-Net model trained as a regression network, second, a 3D U-Net trained as a generator of a generative adversarial network (GAN) and third, a dose transformer model which interprets the dose prediction as a sequence translation task. These models are trained to emulate the result of MC simulations. The dose depositions of a proton minibeam with a diameter of 800μm and an energy of 20-100 MeV inside a simple head phantom calculated by full Geant4 MC simulations are used as a case study for this comparison. The spatial resolution is 0.5 mm. Special attention is put on the evaluation of the generalization performance of the investigatedmodels. Dose predictions with all models are produced in the order of a second on a GPU, the 3D U-Net models being fastest with an average of 130 ms. An investigated 3D U-Net regression model is found to show the strongest performance with overall 61.0 0.5% of all voxels exhibiting a deviation in energy deposition prediction of less than 3% compared to full MC simulations with no spatial deviation allowed. The 3D U-Net models are observed to show better generalization performance for target geometry variations, while the transformer-based model shows better generalization with regard to the protonenergy. This paper reveals that (1) all studied deep learning models are significantly faster than non-machine learning approaches predicting the dose in the order of seconds compared to hours for MC, (2) all models provide reasonable accuracy, and (3) the regression-trained 3D U-Net provides the most accuratepredictions.

Accurate Nonlocal Impact Ionization Models for Conventional and Staircase Avalanche Photodiodes Derived by Full Band Monte Carlo Transport Simulations

We present a procedure to extract the nonlocal impact ionization coefficients in Avalanche Photodiodes (APDs) operating in the linear regime from Full Band Monte Carlo simulations. The Monte Carlo calculations have been calibrated on existing experimental data for GaAs p-i-n APDs with different thickness of the intrinsic region. Inspection of impact ionization generation rate in p-i-n and staircase GaAs APDs led us to identify the limitations of existing nonlocal-history dependent impact ionization models. The introduction of an energy dependent relaxation length for the computation of the effective fields significantly improves the model accuracy in predicting the gain and noise associated to conduction and valence band steps in staircase APDs without additional computational burden. This improved nonlocal-history dependent model is thus a powerful tool to design and optimize APDs with different architectures.