Pith. sign in

REVIEW 3 major objections 5 minor 41 references

Genetic algorithm as a tool for detection setup optimisation: SiFi-CC case study

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A genetic algorithm finds a Compton camera geometry that detects 5 mm range shifts with 2 mm precision.

desk verdict A solid, honest simulation study that gives SiFi-CC a concrete GA-optimized geometry and a 2 mm range-shift resolution at 5e8 protons, but the headline number still needs to survive the neutron background the authors themselves flag. read the letter →

arxiv 2411.18239 v2 pith:ZBNFZCZQ submitted 2024-11-27 physics.med-ph physics.ins-det

Jonas Kasper , Awal Awal , Ronja Hetzel , Magdalena Ko{l}odziej , Katarzyna Rusiecka , Achim Stahl , Ming-Liang Wong , Aleksandra Wrońska This is my paper · ORCID
classification physics.med-phphysics.ins-det
keywords protontherapyprompt-gammaimagingrangeverificationMonteCarlosimulationsComptoncamerageneticalgorithmSiFi-CC
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that a genetic algorithm can systematically find a geometry for the SiFi-CC Compton camera—a device that images prompt gamma rays from a proton beam—that is good enough to detect clinically relevant beam range shifts. The optimized configuration, with 16 fibre layers in the scatterer, 36 in the absorber, and source-scatterer and scatterer-absorber distances of 150 mm and 120 mm, is reported to resolve a 5 mm range shift with 2 mm precision using $5\times 10^8$ protons. This matters because real-time range verification during proton therapy could reduce treatment errors, and the GA approach cuts the parameter search from 7040 full-scan configurations to 100 simulated individuals. The result is based on a detailed Geant4 simulation chain, not on a physical measurement, so the claim is about what the setup can do in a realistic simulated treatment environment.

What carries the argument

The genetic algorithm is the central mechanism: each candidate geometry is encoded as a four-gene individual (SSD, SAD, SL, AL). Its fitness function multiplies four factors: the fraction of prompt-gamma coincidences per impinging proton, the signal-to-background ratio, the quality of event selection, and the reciprocal of the clean image resolution, where that resolution is measured from bootstrapped list-mode MLEM reconstructions as the standard deviation of distal fall-off positions. Each individual's fitness is evaluated through a multi-stage Geant4 simulation chain that models the proton beam time structure, gamma interactions, optical-photon transport in LYSO:Ce fibres, silicon-photomultiplier response, low-level reconstruction, event selection, and image reconstruction; the GA then evolves generations using elitism, gene-pool crossover, and mutation.

What would settle it

Simulate the same optimized geometry with neutron transport in the PMMA phantom and count how many coincident events are neutron-induced; if the correctly-selected Compton fraction falls enough that the distal fall-off resolution at $5\times 10^8$ protons exceeds 2 mm, the central claim is falsified. Alternatively, measure the optimized SiFi-CC in a clinical proton beam with a phantom, and check whether a 5 mm range shift is recovered with 2 mm resolution.

Watch

Extended reading notes

Core claim

The paper's central claim is that the SiFi-CC Compton camera, after genetic-algorithm optimization of four geometric parameters (scatterer layers SL, absorber layers AL, source-to-scatterer distance SSD, scatterer-to-absorber distance SAD), can detect a 5 mm shift in proton beam range with a resolution of 2 mm from prompt-gamma imaging using $5\times 10^8$ impinging protons. The best geometry found has 16 scatterer layers, 36 absorber layers, SSD = 150 mm and SAD = 120 mm, with an imaging sensitivity of $5.58(1)\times 10^{-5}$. In the same simulation, a single $10^8$-proton spot yields only about 5 mm resolution, so the 2 mm figure is reached by combining five spots. The authors also report that the genetic algorithm reduced the number of configurations that had to be evaluated from 7040 to 100, making an otherwise infeasible full parameter scan practical.

Load-bearing premise

The simulation leaves out neutron-induced background and secondary-particle interactions inside the target, so the 2 mm range-shift resolution could be worse in a real clinical beam.

Editorial extensions

If this is right

  • If the 2 mm resolution holds at $5\times 10^8$ protons, a SiFi-CC built to this geometry could verify beam range shifts across multiple treatment spots within clinically relevant timescales.
  • The genetic-algorithm workflow can be reused to optimise other detector parameters, such as materials, pixel sizes, or readout thresholds, without a full parameter scan.
  • Because the single-spot resolution is about 5 mm, per-spot verification is not yet achieved; the reported precision requires combining data from several spots.
  • A hardware coincidence trigger is needed to discard the 89% of single-module events, keeping the data rate near $10^6$ events/s, which existing readout systems can handle.
  • The reported imaging sensitivity of $5.58(1)\times 10^{-5}$ means roughly 5,600 usable events per $10^8$ protons, which supports the multi-spot analysis.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If neutron-induced background reaches the estimated ~20% of coincidences, the correctly selected Compton fraction could drop and the 2 mm resolution might degrade below the clinically useful threshold; the paper does not simulate this, so adding neutron transport to the target is a direct test of the claim.
  • The clean image resolution used in the fitness function is derived from bootstrapped subsets of correctly selected simulated events; realistic event mixing and pile-up may lower the achieved resolution, so beam-test validation is the natural next step.
  • The GA stopped at ten generations without formally meeting its convergence criterion, yet the top five geometries differ by only one step in one parameter; repeated runs with different random seeds could check whether the apparent optimum is global.
  • The same optimisation approach could be transferred to other Compton-camera designs or to imaging detectors whose measurement space is quasi-continuous, where a full system-matrix scan is infeasible.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper presents a genetic-algorithm (GA) based optimisation of the geometry of the SiFi-CC Compton camera for prompt-gamma range verification in proton therapy. Four geometry parameters (SSD, SAD, SL, AL) are optimised with a fitness function that combines detection efficiency, signal-to-background ratio, event-selection quality, and clean-image resolution. Each candidate is evaluated through a Geant4 simulation chain that includes optical photon transport and SiPM response, followed by low-level reconstruction, event selection, and LM-MLEM image reconstruction. The GA stopped after ten generations without meeting its stated convergence criterion. The best configuration (SL=16, AL=36, SSD=150 mm, SAD=120 mm) is then tested on a 5 mm shifted Bragg peak; the authors report a distal fall-off resolution of about 2 mm for 5e8 protons and an imaging sensitivity of 5.58(1)e-5.

Significance. If the result holds, the paper is a useful contribution to prompt-gamma range verification: it demonstrates a systematic, automated framework for geometry optimisation that would be impractical by brute-force parameter scan, and it gives a concrete SiFi-CC configuration with estimated range-shift precision relevant to clinical pencil-beam scanning. The main strength is the end-to-end evaluation of each candidate, including event selection and image reconstruction, rather than a simplified efficiency-only metric. The shifted-phantom test is an appropriate check that the GA result is not merely fitting the training geometry, and the multi-stage simulation chain is described in enough detail to be reproducible. The paper does not provide machine-checked code, but the algorithmic description is sufficiently concrete to be re-implemented.

major comments (3)
  1. [Section 4.1, Fig. 10(b), Conclusion] The central claim that the optimised setup can detect a 5 mm range shift with 2 mm resolution under clinical conditions is not supported by the simulations as described, because secondary-particle and neutron backgrounds are excluded. The authors state that “interactions of secondary particles within the target have not been included” and that the influence of the neutron background “will be investigated”, estimating ≈20% neutron-induced background based on [29]. Figure 10(b) therefore shows event selection and resolution only for gamma-induced coincidences. Since the range-shift resolution depends directly on the event sample composition, the 2 mm value at 5e8 protons is a best-case estimate in an idealised environment. The authors should either include secondary-particle and neutron backgrounds in the simulation and re-evaluate the best geometry, or restrict the conclusions to the gamma-only idealised case and remove “clinical conditions” from the significance statement.
  2. [Section 3.1 with Sections 2.1.3 and Algorithm 1] There is an internal contradiction about convergence. Algorithm 1 defines convergence as three consecutive generation-fitness sums differing by less than 5%, and Section 3.1 reports that the GA terminated at the maximum of ten iterations “since the convergence condition of Algorithm 1 was not met before”. A few paragraphs later the same section states that the 9% to 25% individual fitness spread “shows that the chosen convergence criterion is quite stringent, confirming that the algorithm has successfully converged”. These statements cannot both be true. Because the stopping criterion was not met, the reported best geometry is the best of 100 evaluated individuals, not a certified optimum. The text should be revised to state explicitly that no formal convergence was reached and to avoid claiming convergence.
  3. [Section 3.3, Fig. 10(b), Abstract] The paper equates “capability to detect a 5 mm range shift” with a reported resolution (standard deviation) of about 2 mm, but it does not define a detection criterion or report detection probabilities. With a measured shift of 4.8 mm and a resolution of 2 mm, a single measurement would have a signal-to-noise ratio of about 2.4, which is not by itself “reliable detection” in a clinical sense. The authors should either perform a formal statistical test (e.g., null-distribution based shift detection, ROC analysis, or confidence intervals on the shift) or soften the claim from “reliably detect” to “resolve with a given precision”.
minor comments (5)
  1. [Algorithm 1] The loop condition “while not converged or not 10th generation” should be “while not converged and not at the 10th generation”; as written, the condition is logically unsatisfiable at the maximum generation and does not match the description in Section 2.1.3.
  2. [Section 2.1.2] The word “boostrapping” is a typo for “bootstrapping”.
  3. [Figure 10] The axis labels in the in-text version of Figure 10 are garbled (e.g., “protons 94x10”, “81x10”); the figure itself may be correct in the final PDF, but the text rendering should be fixed.
  4. [Section 2.1.7] The description of the distance-of-closest-approach filter should specify whether the 20 mm distance is measured at the closest approach of the cone to the beam axis or at a fixed depth, and how cones that intersect the beam axis are distinguished from those that do not.
  5. [Section 3.1] The sentence “The variances of their fitness values fall within the 9% to 25% range” should be reworded as “relative standard deviations” if the percentages refer to spreads of fitness values, since a variance is not naturally expressed in percent.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the GA-optimized geometry is validated on an independent shifted-phantom range-shift test, not on the fitness data used during optimization.

full rationale

The paper's derivation chain is self-contained and non-circular. The GA fitness function (Eq. 1) combines detection efficiency, signal-to-background ratio, event-selection quality, and clean image resolution, all evaluated on an unshifted 0 mm Bragg peak simulation for each candidate geometry. The central claim—detection of a 5 mm range shift with 2 mm resolution at 5e8 protons—is tested in Sections 2.3 and 3.3 on a separately simulated 5 mm-shifted phantom, using bootstrapped event subsets and both Monte Carlo truth and experiment-like event selection. No parameter of the range-shift test is fitted to the measured shift; the shift and resolution are outputs. The imaging sensitivity (5.58e-5) is also a derived output from the chosen geometry, not a fitted input. Self-citations such as [16], [17], and [25] provide prior experimental validation of fibre response and of the Geant4 physics-list choice, which are independent external anchors rather than circular premises. The acknowledged omission of neutron and secondary-particle backgrounds (Section 4.1) is a correctness and clinical-validity limitation, not a circular reduction, and the paper explicitly flags it as future work.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. The free parameters are the optimized geometry variables and a few hand-chosen algorithm or analysis thresholds. The main axioms are modeling assumptions about the simulation fidelity and the relevance of the simplified phantom/beam scenario.

free parameters (6)
  • SSD (source-to-scatterer distance) = 150 mm
    Optimized by the genetic algorithm; impacts geometric acceptance and image resolution.
  • SAD (scatterer-to-absorber distance) = 120 mm
    Optimized by the genetic algorithm; affects cone sampling and resolution.
  • SL (scatterer layers) = 16
    Optimized by the genetic algorithm; determines scatterer efficiency and background.
  • AL (absorber layers) = 36
    Optimized by the genetic algorithm; determines absorption efficiency and readout count.
  • GA hyperparameters (population 10, generations 10, mutation rate 3+) = See Section 2.1.2-2.1.6
    Chosen by hand based on computing constraints and prior practice; affect search convergence, not the physics directly.
  • Event selection distance-of-closest-approach cut = 20 mm
    Ad hoc threshold in the event selection; affects the balance of signal and background in reconstructed images.
assumptions (4)
  • domain assumption Geant4 physics list QGSP_BIC_HP_EMZ accurately models prompt gamma production and transport in the simulated energy range.
    Relied on throughout Section 2.1.7; earlier validation only for a single spectrum, not for all geometries.
  • domain assumption The PMMA phantom and 130 MeV proton beam are representative of clinical proton therapy conditions.
    Clinical targets vary in composition and geometry; the optimized geometry may not transfer to other beam energies or tissue types.
  • domain assumption The 'clean image resolution' sigmaCleanImg, derived from bootstrapped subsets of 300 correctly selected events, is a valid proxy for the actual range shift detection resolution.
    Used in the fitness function (Eq. 1) and later compared to the full range-shift resolution; the proxy ignores systematic effects from event selection background.
  • standard math The standard deviation of distal fall-off positions across bootstrap subsets is an unbiased estimator of the statistical resolution.
    Applied in Section 2.1.7 and Section 3.3 to quantify resolution; assumes the bootstrap subsets are independent and representative.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Genetic algorithm as a tool for detection setup optimisation: SiFi-CC case study." pith.science (2026). https://pith.science/paper/ZBNFZCZQ

@misc{pith2026241118239,
  author       = {Pith},
  title        = {Pith review of: Genetic algorithm as a tool for detection setup optimisation: SiFi-CC case study},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZBNFZCZQ}},
  note         = {Machine review of arXiv:2411.18239}
}
read the original abstract

Objective: Proton therapy is a precision-focused cancer treatment where accurate proton beam range monitoring is critical to ensure effective dose delivery. This can be achieved by prompt gamma detection with a Compton camera like the SiFi-CC. This study aims to show the feasibility of optimising the geometry of SiFi-CC Compton camera for verification of dose distribution via prompt gamma detection using a genetic algorithm (GA). Approach: The SiFi-CC key geometric parameters for optimisation with the GA are the source-to-scatterer and scatterer-to-absorber distances, and the module thicknesses. The optimisation process was conducted with a software framework based on the Geant4 toolkit, which included detailed and realistic modelling of gamma interactions, detector response, and further steps such as event selection and image reconstruction. The performance of each individual configuration was evaluated using a fitness function incorporating factors related to gamma detection efficiency and image resolution. Results: The GA-optimised SiFi-CC configuration demonstrated the capability to detect a 5 mm proton beam range shift with a 2 mm resolution using 5e8 protons. The best-performing geometry, with 16 fibre layers in the scatterer, 36 layers in the absorber, source-to-scatterer distance 150 mm and scatterer-to-absorber distance 120 mm, has an imaging sensitivity of 5.58(1)e-5. Significance: This study demonstrates that the SiFi-CC setup, optimised through a GA, can reliably detect clinically relevant proton beam range shifts, improving real-time range verification accuracy in proton therapy. The presented implementation of a GA is a systematic and feasible way of searching for a SiFi-CC geometry that shows the best performance.

Figures

Figures reproduced from arXiv: 2411.18239 by the authors.

Figure 1
Figure 1. Perspective view of an example Compton camera: the SiFi-CC setup, with the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. For this selection example, the individuals (in descending fitness score) up to D [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. As an example, only four parents are considered for this crossover step. The genes [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Example of mutation: three of the four individuals have a single gene selected for [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Steps needed to evaluate an individual in the course of GA. The multi-stage Monte [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Time ion of the simulated PGs reflecting the time structure of the clinical proton [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: The distal fall-off position in a PG depth profile as the inflection point (red dot) of [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: (a) Fitness of the individuals evaluated by GA, intervals of 10 on the horizontal axis [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Number of evaluations and average fitness for tested individuals, represented by [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Study of distal fall-off (inflection point of a sigmoid fit) position resolution as a [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 27 canonical work pages

  1. [29]

    Noise evaluation of Comp- ton camera imaging for proton therapy

    Ortega PG, Torres-Espallardo I, Cerutti F, et al. Noise evaluation of Comp- ton camera imaging for proton therapy. Physics in Medicine and Biology 2015;60:1845–63. doi: 10.1088/0031-9155/60/5/1845. 19

    work page doi:10.1088/0031-9155/60/5/1845 2015

  2. [1]

    First clinical application of a prompt gamma based in vivo proton range verification using a knife-edge slit cam- era

    Richter C, Pausch G, Barczyk S, et al. First clinical application of a prompt gamma based in vivo proton range verification using a knife-edge slit cam- era. Radiotherapy and Oncology 2016;118:S89–S90. doi: 10.1016/S0167- 8140(16)30184-0

    doi:10.1016/s0167- 2016

  3. [2]

    Detectability of Anatomical Changes With Prompt-Gamma Imaging: First Systematic Evaluation of Clinical Application During Prostate-Cancer Proton Therapy

    Berthold J, Pietsch J, Piplack N, et al. Detectability of Anatomical Changes With Prompt-Gamma Imaging: First Systematic Evaluation of Clinical Application During Prostate-Cancer Proton Therapy. International Jour- nal of Radiation Oncology Biology Physics 3 2023;117:718–29. doi: 10. 1016/j.ijrobp.2023.05.002

    work page 2023

  4. [3]

    Prompt Gamma Imaging for In Vivo Range Verification of Pencil Beam Scanning Proton Therapy

    Xie Y, Bentefour EH, Janssens G, et al. Prompt Gamma Imaging for In Vivo Range Verification of Pencil Beam Scanning Proton Therapy. Inter- national Journal of Radiation Oncology •Biology•Physics 2017;99:210–8. doi: https://doi.org/10.1016/j.ijrobp.2017.04.027

    doi:10.1016/j.ijrobp.2017.04.027 2017

  5. [4]

    Prompt gamma imaging for the iden- tification of regional proton range deviations due to anatomic change in a heterogeneous region

    Xie Y, Petzoldt J, Janssens G, et al. Prompt gamma imaging for the iden- tification of regional proton range deviations due to anatomic change in a heterogeneous region. British Journal of Radiology 1116 2020;93:20190619. doi: DOI10.1259/bjr.20190619. url: https://academic.oup.com/bjr/ article/93/1116/20190619/7452133. 17

    work page 2020

  6. [5]

    A full-scale clinical prototype for proton range verification using prompt gamma-ray spectroscopy

    Hueso-Gonz´ alez F, Rabe M, Ruggieri TA, Bortfeld T, and Verburg JM. A full-scale clinical prototype for proton range verification using prompt gamma-ray spectroscopy. Physics in Medicine and Biology 2018;63 18:185019. doi: 10.1088/1361-6560/aad513

    work page doi:10.1088/1361-6560/aad513 2018

  7. [6]

    Range verification by means of prompt- gamma detection in particle therapy

    Wro´ nska A and Dauvergne D. Range verification by means of prompt- gamma detection in particle therapy. In: Radiation Detection Systems, vol. 2. Ed. by Iniewski K and Iwa´ nczyk J. Devices, Circuits, and Systems. CRC Press/Routledge, 2021. doi: 10.1201/9781003218364-6

    work page doi:10.1201/9781003218364-6 2021

  8. [7]

    Prompt-gamma imaging in particle therapy

    Pinto M. Prompt-gamma imaging in particle therapy. European Physical Journal Plus 10 2024;139:844. doi: 10.1140/epjp/s13360-024-05664-4

    doi:10.1140/epjp/s13360-024-05664-4 2024

Show all 41 references
  1. [8]

    Detection systems for range mon- itoring in proton therapy: Needs and challenges

    Pausch G, Berthold J, Roemer K, et al. Detection systems for range mon- itoring in proton therapy: Needs and challenges. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, De- tectors and Associated Equipment 2018;954. doi: 10.1016/j.ni...

    2018 doi

  2. [9]

    The effects of Compton camera data acquisition and readout timing on PG imaging for proton range verification

    Polf JC, Maggi P, Panthi R, Peterson S, Mackin D, and Beddar S. The effects of Compton camera data acquisition and readout timing on PG imaging for proton range verification. IEEE Transactions on Radiation and Plasma Medical Sciences 2021:1–1. doi: 10.1109/TRPMS.2021.3057341

    2021

  3. [10]

    Tests of a Compton imag- ing prototype in a monoenergetic 4.44 MeV photon field—a benchmark setup for prompt gamma-ray imaging devices

    Golnik C, Bemmerer D, Enghardt W, et al. Tests of a Compton imag- ing prototype in a monoenergetic 4.44 MeV photon field—a benchmark setup for prompt gamma-ray imaging devices. Journal of Instrumentation 2016;11:P06009–P06009. doi: 10.1088/1748-0221/11/06/p06009

    2016 doi

  4. [11]

    Characterization of a Compton camera setup with monolithic LaBr3(Ce) absorber and segmented GAGG scatter detectors

    Liprandi S, Takyu S, Aldawood S, et al. Characterization of a Compton camera setup with monolithic LaBr3(Ce) absorber and segmented GAGG scatter detectors. 2017 IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC) 2017:1–4. doi: 10.1109/NSSMIC.2017. 8533134

    2017 doi

  5. [12]

    Development of a Compton camera for medical applications based on silicon strip and scin- tillation detectors

    Krimmer J, Abellan Beteta C, Cachemiche JP, et al. Development of a Compton camera for medical applications based on silicon strip and scin- tillation detectors. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associate...

    2014 doi

  6. [13]

    Optimization of collimator designs for real-time proton range verification by measuring prompt gamma rays

    Lopes PC, Pinto M, Simoes H, et al. Optimization of collimator designs for real-time proton range verification by measuring prompt gamma rays. In: IEEE, 2012:3864–70. doi: 10.1109/NSSMIC.2012.6551886 . url: http: //ieeexplore.ieee.org/document/6551886/

    2012

  7. [14]

    Design optimisation of a TOF- based collimated camera prototype for online hadrontherapy monitoring

    Pinto M, Dauvergne D, Freud N, et al. Design optimisation of a TOF- based collimated camera prototype for online hadrontherapy monitoring. Physics in Medicine and Biology 24 2014;59:7653–74. doi: 10.1088/0031- 9155/59/24/7653. url: http://stacks.iop.org/0031-9155/59/i=24/ a=76...

    2014 doi

  8. [15]

    Distal edge determination pre- cision for a multi-slat promptgamma camera: a comprehensive simula- tion and optimization of the detection system

    Morozov A, Sim˜ oes H, and Crespo P. Distal edge determination pre- cision for a multi-slat promptgamma camera: a comprehensive simula- tion and optimization of the detection system. 2021. arXiv: 2009.14154 [physics.med-ph]. 18

    2021 arXiv

  9. [16]

    Optimisation of the SiFi-CC Compton camera for range verifica- tion in proton therapy through a genetic algorithm

    Kasper J. Optimisation of the SiFi-CC Compton camera for range verifica- tion in proton therapy through a genetic algorithm. Ph.D. thesis. R WTH Aachen University, 2022. doi: 10.18154/RWTH-2022-11279

    2022 doi

  10. [17]

    The SiFi-CC project – Feasibility study of a scintillation-fiber-based Compton camera for proton therapy monitoring

    Kasper J, Rusiecka K, Hetzel R, et al. The SiFi-CC project – Feasibility study of a scintillation-fiber-based Compton camera for proton therapy monitoring. Physica Medica 2020;76:317–25. doi: https://doi.org/10. 1016/j.ejmp.2020.07.013

    2020

  11. [18]

    Adaptation in natural and artificial systems

    Holland J. Adaptation in natural and artificial systems. MIT Press, 1975

    1975

  12. [19]

    An Overview of Genetic Algorithms: Part 1, Fundamentals

    Beasley D, Bull DR, and Martin RR. An Overview of Genetic Algorithms: Part 1, Fundamentals. University Computing 1993;15:58–69

    1993

  13. [20]

    Chapter 6 - Genetic Algorithms

    Yang XS. Chapter 6 - Genetic Algorithms. In: Nature-Inspired Optimiza- tion Algorithms (Second Edition). Ed. by Yang XS. Second Edition. Aca- demic Press, 2021:91–100. doi: https://doi.org/10.1016/B978-0-12- 821986-7.00013-5

    2021 doi

  14. [21]

    Optimization of a neutrino beam for the study of CP violation with the LENA and JUNO detector

    Meloni M. Optimization of a neutrino beam for the study of CP violation with the LENA and JUNO detector. Dissertation. Aachen: R WTH Aachen University, 2016:85 Seite. url: https://publications.rwth- aachen. de/record/679982

    2016

  15. [22]

    Gene Pool Recombination in Genetic Algorithms

    M¨ uhlenbein H. Gene Pool Recombination in Genetic Algorithms. In:Meta- Heuristics. Ed. by Osman I and Kelly J. Springer, 1996:53–65. doi: 10. 1007/978-1-4613-1361-8_4

    1996

  16. [23]

    GEANT4–a simulation toolkit

    Agostinelli S et al. GEANT4–a simulation toolkit. Nucl. Instrum. Meth. A 2003;506:250–303. doi: 10.1016/S0168-9002(03)01368-8

    2003 doi

  17. [24]

    https://geant4- userdoc.web.cern.ch/ UsersGuides/PhysicsListGuide/fo/PhysicsListGuide.pdf

    Geant4: Physics List Guide. https://geant4- userdoc.web.cern.ch/ UsersGuides/PhysicsListGuide/fo/PhysicsListGuide.pdf . Accessed: 2024-10-29

    2024

  18. [25]

    Prompt-gamma emission in GEANT4 revisited and confronted with experiment

    Wro´ nska A, Kasper J, Anees Ahmed A, et al. Prompt-gamma emission in GEANT4 revisited and confronted with experiment. Physica Medica 2021;88:250–61. doi: 10.1016/j.ejmp.2021.07.018

    2021 doi

  19. [26]

    MetaLaser. Ce:LYSO. url: https : / / www . meta - laser . com / Content / upload/PDF/20199763/Ce- LYSO- Scintillator- Metalaser.pdf?rnd= 276 (visited on 03/10/2022)

    2022

  20. [27]

    The effects of Doppler broad- ening and detector resolution on the performance of three-stage Compton cameras

    MacKin D, Polf J, Peterson S, and Beddar S. The effects of Doppler broad- ening and detector resolution on the performance of three-stage Compton cameras. Medical Physics 1 2013;40. doi: 10.1118/1.4767756

    2013 doi

  21. [28]

    List-mode maximum likelihood reconstruction of Compton scatter camera images in nuclear medicine

    Wilderman S, Clinthorne N, Fessler J, and Rogers W. List-mode maximum likelihood reconstruction of Compton scatter camera images in nuclear medicine. In: 1998 IEEE Nuclear Science Symposium Conference Record. 1998 IEEE Nuclear Science Symposium and Medical Imaging Conference (...

    1998 doi

  22. [30]

    Proton range verification with MACACO II Compton camera enhanced by a neural network for event selection

    Mu˜ noz E, Ros Garc ´ ıa A, Borja-Lloret M, et al. Proton range verification with MACACO II Compton camera enhanced by a neural network for event selection. Scientific Reports 2021;11:9325. doi: 10.1038/s41598- 021-88812-5

    2021 doi

  23. [31]

    Machine learning-based event recognition in SiFi Compton camera imaging for proton therapy monitoring

    Kozani MK and Magiera A. Machine learning-based event recognition in SiFi Compton camera imaging for proton therapy monitoring. Physics in Medicine and Biology 15 2022;67. doi: 10.1088/1361-6560/ac71f2

    2022 doi

  24. [32]

    Towards machine learning aided real-time range imaging in proton therapy

    Lerendegui-Marco J, Balibrea-Correa J, Babiano-Su´ arez V, Ladarescu I, and Domingo-Pardo C. Towards machine learning aided real-time range imaging in proton therapy. Scientific Reports 1 2022;12. doi: 10.1038/ s41598-022-06126-6

    2022

  25. [33]

    Promising Hyperparameter Con- figurations for Deep Fully Connected Neural Networks to Improve Image Reconstruction in Proton Radiotherapy

    York SA, Ali AM, Lashbrooke DC, et al. Promising Hyperparameter Con- figurations for Deep Fully Connected Neural Networks to Improve Image Reconstruction in Proton Radiotherapy. In: 2021 IEEE International Con- ference on Big Data (Big Data). 2021:5648–57. doi: 10.1109/BigData...

    2021

  26. [34]

    Applications of Machine Learning to Improve the Clinical Viability of Compton Camera Based in vivo Range Verification in Proton Radiotherapy

    Polf JC, Barajas CA, Peterson SW, et al. Applications of Machine Learning to Improve the Clinical Viability of Compton Camera Based in vivo Range Verification in Proton Radiotherapy. Frontiers in Physics 2022;10. doi: 10.3389/fphy.2022.838273

    2022

  27. [35]

    Deep residual fully connected neural network classification of Compton camera based prompt gamma imaging for proton radiotherapy

    Barajas CA, Polf JC, and Gobbert MK. Deep residual fully connected neural network classification of Compton camera based prompt gamma imaging for proton radiotherapy. Frontiers in Physics 2023;11. doi: 10. 3389/fphy.2023.903929

    2023

  28. [36]

    Comparison of readout systems for high-rate silicon photomultiplier applications

    Wong M, Ko lodziej M, Briggl K, et al. Comparison of readout systems for high-rate silicon photomultiplier applications. Journal of Instrumentation 2024;19:P01019. doi: 10.1088/1748-0221/19/01/P01019

    2024 doi

  29. [37]

    Capability of MLEM and OE to Detect Range Shifts with a Compton Camera in Particle Therapy

    Kohlhase N, Wegener T, Schaar M, et al. Capability of MLEM and OE to Detect Range Shifts with a Compton Camera in Particle Therapy. IEEE Transactions on Radiation and Plasma Medical Sciences 2020;4:233–42. doi: 10.1109/TRPMS.2019.2937675

    2020

  30. [38]

    Monitoring Ion Beam Therapy With a Compton Camera: Simulation Studies of the Clinical Feasibility

    Fontana M, Dauvergne D, Freud N, et al. Monitoring Ion Beam Therapy With a Compton Camera: Simulation Studies of the Clinical Feasibility. IEEE Transactions on Radiation and Plasma Medical Sciences 2019;PP:1–

    2019

  31. [39]

    doi: 10.1109/TRPMS.2019.2933985

    2019

  32. [40]

    3D prompt gamma imaging for pro- ton beam range verification

    Draeger E, Mackin D, Peterson S, et al. 3D prompt gamma imaging for pro- ton beam range verification. Physics in Medicine and Biology 2018;63:035019. doi: 10.1088/1361-6560/aaa203

    2018 doi

  33. [41]

    Analytical modeling and Monte Carlo simulations of multi-parallel slit and knife-edge slit prompt gamma cameras

    Huisman BFB, Mu˜ noz E, Dauvergne D, L´ etang JM, Sarrut D, and Testa´E. Analytical modeling and Monte Carlo simulations of multi-parallel slit and knife-edge slit prompt gamma cameras. Physics in Medicine and Biology 11 2023;68:115009. doi: 10 . 1088 / 1361 - 6560 / acd237. u...

    2023 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.