pith. sign in

arxiv: 2606.12602 · v1 · pith:GVKGNN5Inew · submitted 2026-06-10 · ⚛️ physics.med-ph

Beam modelling of Hitachi PROBEAT proton therapy system for a GPU-based Fast Monte Carlo dose engine

Qianxia Wang , Poenisch Falk , Yao Zhao , Xueming Bai , Roelf Slopsema , Kirk Jon Luca , Thomas J Whitaker , Yun Hu
show 3 more authors
Uwe Titt Radhe Mohan Pablo Yepes
This is my paper

Pith reviewed 2026-06-27 07:17 UTC · model grok-4.3

classification ⚛️ physics.med-ph
keywords proton therapybeam modelingdouble-Gaussianphase space filesMonte Carlodose engineHitachi PROBEAT
0
0 comments X

The pith

Upgrading beam model to double-Gaussian distributions yields more accurate phase space files for proton Monte Carlo engine.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors upgrade their beam model for the Hitachi PROBEAT proton therapy system by applying double-Gaussian distributions to both the spatial distribution and the opening angle distribution of particles. This upgrade is meant to produce more accurate phase space files as input for their GPU-based fast Monte Carlo dose engine. A reader would care because this independent engine supports patient QA verification and research without depending on the clinic treatment planning system. The change targets improved performance at the expanded proton center through better modeling of particle properties.

Core claim

Upgrading the beam model with double-Gaussian distributions for both spatial and opening angle distributions of particles obtains more accurate phase space files for the GPU-based Monte Carlo dose engine.

What carries the argument

Double-Gaussian distributions for spatial and opening angle distributions of particles, replacing prior distributions in the beam model to generate phase space files.

If this is right

  • The phase space files become more accurate inputs for dose calculations.
  • The independent dose engine gains reliability for patient QA verification.
  • New functions can be implemented more effectively in research applications.
  • Overall performance of the dose engine improves for the expanded proton center.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the upgrade works, clinics could adopt similar beam modeling changes to enhance independent verification tools.
  • Testing the new model against measured dose data in phantoms would directly check the accuracy gain.
  • Similar double-Gaussian approaches might apply to other particle therapy systems beyond Hitachi PROBEAT.

Load-bearing premise

Switching to double-Gaussian distributions for spatial and angular particle distributions produces more accurate phase space files than the prior model.

What would settle it

Comparison of dose calculations using the upgraded phase space files versus the old ones against actual measurements, showing equivalent or lower accuracy with the double-Gaussian model.

Figures

Figures reproduced from arXiv: 2606.12602 by Kirk Jon Luca, Pablo Yepes, Poenisch Falk, Qianxia Wang, Radhe Mohan, Roelf Slopsema, Thomas J Whitaker, Uwe Titt, Xueming Bai, Yao Zhao, Yun Hu.

Figure 1
Figure 1. Figure 1: From left to right are intergral depth dose (IDD) for 70.2, 79.2, 89.2, 98.4, 107.0, 115.1, 124.7, 134.3, 143.5, 152.4, 160.3, 168.0, 175.5, 183.7, 192.1, 200.3, 208.3, 216.1, 223.7 and 228.7 MeV/A, respectively. Black solid lines are for FDC simulations and red dot-lines are for measurements [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Numbers of protons per MU for different energies. The red triangles are values derived from measurement for the 24 energies: 70.2, 73.9, 79.2, 84.3, 89.2, 93.9, 98.4, 102.7, 107.0, 115.1, 124.7, 134.3, 143.5, 152.4, 160.3, 168.0, 175.5, 183.7, 192.1, 200.3, 208.3, 216.1, 223.7 and 228.7 MeV/A. The black spots are interpolation values for the rest of the available energies [PITH_FULL_IMAGE:figures/full_fig… view at source ↗
read the original abstract

Background: An in-house dose engine independent of clinic TPS is not only a reliable tool for patient QA verification. More importantly, it plays vital role in cutting-edge research due to its flexibility in implementing new functions. In this study, we upgraded our existing beam model with using double-Gaussian distributions for both spatial and opening angle distributions of particles to obtain more accurate phase space files. It is expected to potentially improve the performance of this independent dose engine in both clinic and research at the expanded MD Anderson proton center.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The manuscript describes an upgrade to the beam model for the Hitachi PROBEAT proton therapy system, replacing prior distributions with double-Gaussian forms for both the spatial and opening-angle distributions of particles in order to generate more accurate phase-space files for an in-house GPU-based fast Monte Carlo dose engine intended for patient QA and research at the expanded MD Anderson proton center.

Significance. If the double-Gaussian model were shown to measurably improve dosimetric accuracy, the work would strengthen the reliability of independent, TPS-independent dose engines in proton therapy; the GPU implementation addresses a practical need for rapid calculations. However, the absence of any validation data means the claimed accuracy gain remains untested.

major comments (1)
  1. [Abstract] Abstract: the central claim that the double-Gaussian upgrade 'obtains more accurate phase space files' is unsupported; the text only states that the change is 'expected to potentially improve' performance and supplies no comparison metrics (gamma analysis, depth-dose differences, lateral profiles, or error bars) against measurements or the prior single-Gaussian model.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the detailed review and constructive comment. We agree that the abstract wording requires correction to avoid overstating results, and we will revise the manuscript accordingly. Our point-by-point response follows.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the double-Gaussian upgrade 'obtains more accurate phase space files' is unsupported; the text only states that the change is 'expected to potentially improve' performance and supplies no comparison metrics (gamma analysis, depth-dose differences, lateral profiles, or error bars) against measurements or the prior single-Gaussian model.

    Authors: We acknowledge the referee's observation. The manuscript body accurately describes the upgrade as one that 'is expected to potentially improve' performance, while the abstract uses stronger language claiming that it 'obtains more accurate phase space files.' This discrepancy was an inadvertent drafting inconsistency. We will revise the abstract to match the body text, removing any unsupported claim of demonstrated accuracy improvement. This manuscript focuses on the technical implementation of the double-Gaussian beam model; no comparative validation metrics against measurements or the prior single-Gaussian model are included here, as such dosimetric validation is reserved for a separate study. revision: yes

Circularity Check

0 steps flagged

No circularity identified in the beam modeling description

full rationale

The manuscript describes upgrading an existing beam model by adopting double-Gaussian distributions for spatial and opening-angle particle distributions in order to obtain more accurate phase space files for a GPU-based Monte Carlo engine. No equations, parameter-fitting procedures, self-citations, or uniqueness theorems are supplied that would reduce the accuracy claim to a quantity defined by the authors' own prior inputs or fits. The central assertion remains an unvalidated modeling choice rather than a derivation that collapses by construction, leaving the paper self-contained against external benchmarks for the purpose of circularity analysis.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; no explicit free parameters, axioms, or invented entities can be extracted. The double-Gaussian choice itself functions as an unstated modeling assumption whose justification is not supplied.

pith-pipeline@v0.9.1-grok · 5645 in / 1048 out tokens · 13740 ms · 2026-06-27T07:17:51.667596+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

28 extracted references · 24 canonical work pages

  1. [1]

    Proton therapy – Present and future

    Mohan R, Grosshans D. Proton therapy – Present and future. Adv Drug Deliv Rev. 2017;109:26-44. doi:10.1016/j.addr.2016.11.006

    work page doi:10.1016/j.addr.2016.11.006 2017

  2. [2]

    Accessed April 12, 2026

    Why Choose MD Anderson Proton Therapy? UT MD Anderson. Accessed April 12, 2026. https://www.mdanderson.org/patients-family/diagnosis-treatment/care-centers-clinics/proton-therapy- center/why-choose-md-anderson-proton-therapy.html

    2026

  3. [3]

    Accessed April 12, 2026

    PTCOG - Facilities in Operation (public). Accessed April 12, 2026. https://www.ptcog.site/index.php/facilities-in-operation-public

    2026

  4. [4]

    Monte Carlo fast dose calculator for proton radiotherapy: application to a voxelized geometry representing a patient with prostate cancer

    Yepes P, Randeniya S, Taddei PJ, Newhauser WD. Monte Carlo fast dose calculator for proton radiotherapy: application to a voxelized geometry representing a patient with prostate cancer. Phys Med Biol. 2008;54(1):N21. doi:10.1088/0031-9155/54/1/N03

    work page doi:10.1088/0031-9155/54/1/n03 2008

  5. [5]

    A GPU implementation of a track-repeating algorithm for proton radiotherapy dose calculations

    Yepes PP, Mirkovic D, Taddei PJ. A GPU implementation of a track-repeating algorithm for proton radiotherapy dose calculations. Phys Med Biol. 2010;55(23):7107. doi:10.1088/0031-9155/55/23/S11

    work page doi:10.1088/0031-9155/55/23/s11 2010

  6. [6]

    Linear energy transfer incorporated intensity modulated proton therapy optimization

    Cao W, Khabazian A, Yepes P, et al. Linear energy transfer incorporated intensity modulated proton therapy optimization. Phys Med Biol. 2017;63(1):015013. doi:10.1088/1361-6560/aa9a2e

    work page doi:10.1088/1361-6560/aa9a2e 2017

  7. [7]

    An empirical model of proton RBE based on the linear correlation between x-ray and proton radiosensitivity

    Flint DB, Ruff CE, Bright SJ, et al. An empirical model of proton RBE based on the linear correlation between x-ray and proton radiosensitivity. Med Phys. 2022;49(9):6221-6236. doi:10.1002/mp.15850 9

    work page doi:10.1002/mp.15850 2022

  8. [8]

    Fixed- versus Variable-RBE Computations for Intensity Modulated Proton Therapy

    Yepes P, Adair A, Frank SJ, et al. Fixed- versus Variable-RBE Computations for Intensity Modulated Proton Therapy. Adv Radiat Oncol. 2019;4(1):156-167. doi:10.1016/j.adro.2018.08.020

    work page doi:10.1016/j.adro.2018.08.020 2019

  9. [9]

    Intensity modulated proton arc therapy via geometry‐based energy selection for ependymoma

    Cao W, Li Y, Zhang X, et al. Intensity modulated proton arc therapy via geometry‐based energy selection for ependymoma. J Appl Clin Med Phys. 2023;24(7):e13954. doi:10.1002/acm2.13954

    work page doi:10.1002/acm2.13954 2023

  10. [10]

    A track repeating algorithm for intensity modulated carbon ion therapy

    Wang Q, Adair A, Deng Y, et al. A track repeating algorithm for intensity modulated carbon ion therapy. Phys Med Biol. 2019;64(9):095026. doi:10.1088/1361-6560/ab10d0

    work page doi:10.1088/1361-6560/ab10d0 2019

  11. [11]

    Optimization of FLASH proton beams using a track-repeating algorithm

    Wang Q, Titt U, Mohan R, et al. Optimization of FLASH proton beams using a track-repeating algorithm. Med Phys. 2022;49(10):6684-6698. doi:10.1002/mp.15849

    work page doi:10.1002/mp.15849 2022

  12. [12]

    Validation of a track-repeating algorithm versus measurements in water for proton scanning beams

    Yepes PP, Guan F, Kerr M, et al. Validation of a track-repeating algorithm versus measurements in water for proton scanning beams. Biomed Phys Eng Express. 2016;2(3):037002. doi:10.1088/2057-1976/2/3/037002

    work page doi:10.1088/2057-1976/2/3/037002 2016

  13. [13]

    Assessment of the accuracy of an MCNPX-based Monte Carlo simulation model for predicting three-dimensional absorbed dose distributions

    Titt U, Sahoo N, Ding X, et al. Assessment of the accuracy of an MCNPX-based Monte Carlo simulation model for predicting three-dimensional absorbed dose distributions. Phys Med Biol. 2008;53(16):4455. doi:10.1088/0031-9155/53/16/016

    work page doi:10.1088/0031-9155/53/16/016 2008

  14. [14]

    Validation of the fast dose calculator for Shanghai Proton and Heavy Ion Center

    Wang Q, Schlegel N, Moyers M, et al. Validation of the fast dose calculator for Shanghai Proton and Heavy Ion Center. Biomed Phys Eng Express. 2018;4(6):065007. doi:10.1088/2057-1976/aae039

    work page doi:10.1088/2057-1976/aae039 2018

  15. [15]

    Automatic phase space generation for Monte Carlo calculations of intensity modulated particle therapy

    Wang Q, Zhu C, Bai X, et al. Automatic phase space generation for Monte Carlo calculations of intensity modulated particle therapy. Biomed Phys Eng Express. 2020;6(2):025001. doi:10.1088/2057-1976/ab7152

    work page doi:10.1088/2057-1976/ab7152 2020

  16. [16]

    Beyond Gaussians: a study of single-spot modeling for scanning proton dose calculation

    Li Y, Zhu RX, Sahoo N, Anand A, Zhang X. Beyond Gaussians: a study of single-spot modeling for scanning proton dose calculation. Phys Med Biol. 2012;57(4):983. doi:10.1088/0031-9155/57/4/983

    work page doi:10.1088/0031-9155/57/4/983 2012

  17. [17]

    Producing a Beam Model of the Varian ProBeam Proton Therapy System using TOPAS Monte Carlo Toolkit

    Rahman M, Bruza P, Lin Y, Gladstone DJ, Pogue BW, Zhang R. Producing a Beam Model of the Varian ProBeam Proton Therapy System using TOPAS Monte Carlo Toolkit. Med Phys. 2020;47(12):6500-6508. doi:10.1002/mp.14532

    work page doi:10.1002/mp.14532 2020

  18. [18]

    A technique for the quantitative evaluation of dose distributions

    Low DA, Harms WB, Mutic S, Purdy JA. A technique for the quantitative evaluation of dose distributions. Med Phys. 1998;25(5):656-661. doi:10.1118/1.598248

    work page doi:10.1118/1.598248 1998

  19. [19]

    Treatment planning of scanned proton beams in RayStation

    Janson M, Glimelius L, Fredriksson A, Traneus E, Engwall E. Treatment planning of scanned proton beams in RayStation. Med Dosim. 2024;49(1):2-12. doi:10.1016/j.meddos.2023.10.009

    work page doi:10.1016/j.meddos.2023.10.009 2024

  20. [20]

    Clinical commissioning of intensity-modulated proton therapy systems: Report of AAPM Task Group 185

    Farr JB, Moyers MF, Allgower CE, et al. Clinical commissioning of intensity-modulated proton therapy systems: Report of AAPM Task Group 185. Med Phys. 2021;48(1):e1-e30. doi:10.1002/mp.14546

    work page doi:10.1002/mp.14546 2021

  21. [21]

    Experimental characterization and physical modelling of the dose distribution of scanned proton pencil beams

    Pedroni E, Scheib S, Böhringer T, et al. Experimental characterization and physical modelling of the dose distribution of scanned proton pencil beams. Phys Med Biol. 2005;50(3):541. doi:10.1088/0031-9155/50/3/011

    work page doi:10.1088/0031-9155/50/3/011 2005

  22. [22]

    Commissioning dose computation models for spot scanning proton beams in water for a commercially available treatment planning system

    Zhu XR, Poenisch F, Lii M, et al. Commissioning dose computation models for spot scanning proton beams in water for a commercially available treatment planning system. Med Phys. 2013;40(4):041723. doi:10.1118/1.4798229

    work page doi:10.1118/1.4798229 2013

  23. [23]

    Validating a double Gaussian source model for small proton fields in a commercial Monte-Carlo dose calculation engine

    Kugel F, Wulff J, Bäumer C, et al. Validating a double Gaussian source model for small proton fields in a commercial Monte-Carlo dose calculation engine. Z Für Med Phys. 2023;33(4):529-541. doi:10.1016/j.zemedi.2022.11.011 10

    work page doi:10.1016/j.zemedi.2022.11.011 2023

  24. [24]

    Accessed May 4, 2026

    Nuclear halo measurements for accurate prediction of field size factor in a Varian ProBeam proton PBS system - Harms - 2020 - Journal of Applied Clinical Medical Physics - Wiley Online Library. Accessed May 4, 2026. https://aapm.onlinelibrary.wiley.com/doi/full/10.1002/acm2.12783

    work page doi:10.1002/acm2.12783 2020

  25. [25]

    Comprehensive investigation of lateral dose profile and output factor measurements in small proton fields from different delivery techniques

    Kretschmer J, Brodbek L, Behrends C, et al. Comprehensive investigation of lateral dose profile and output factor measurements in small proton fields from different delivery techniques. Med Phys. 2023;50(7):4546-4561. doi:10.1002/mp.16357

    work page doi:10.1002/mp.16357 2023

  26. [26]

    Development and validation of an automatic commissioning tool for the Monte Carlo dose engine in myQA iON

    Cohilis M, Hong L, Janssens G, et al. Development and validation of an automatic commissioning tool for the Monte Carlo dose engine in myQA iON. Phys Medica PM Int J Devoted Appl Phys Med Biol Off J Ital Assoc Biomed Phys AIFB. 2022;95:1-8

    2022

  27. [27]

    Automation of Monte Carlo‐based treatment plan verification for proton therapy

    Kaluarachchi M, Moskvin V, Pirlepesov F, Wilson LJ, Xie F, Faught AM. Automation of Monte Carlo‐based treatment plan verification for proton therapy. J Appl Clin Med Phys. Published online 2020. doi:10.1002/acm2.12923

    work page doi:10.1002/acm2.12923 2020

  28. [28]

    Monte Carlo simulation‐based patient‐specific QA using machine log files for line‐scanning proton Figure 1. From left to right are intergral depth dose (IDD) for 70.2, 79.2, 89.2, 98.4, 107.0, 115.1, 124.7, 134.3, 143.5, 152.4, 160.3, 168.0, 175.5, 183.7, 1 92.1, 200.3, 208.3, 216.1, 223.7 and 228.7 M eV/A, respectively. Black solid lines are for FDC simu...