Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks

Helge Egil Seime Pettersen; Max Aehle; Johan Alme; Gergely Gábor Barnaföldi; Vyacheslav Borshchov; Anthony van den Brink; Mamdouh Chaar; Viljar Eikeland; Grigory Feofilov; Christoph Garth; Nicolas R. Gauger; Georgi Genov; Ola Grøttvik; Håvard Helstrup; Sergey Igolkin; Ralf Keidel; Chinorat Kobdaj; Tobias Kortus; Viktor Leonhardt; Shruti Mehendale; Raju Ningappa Mulawade; Odd Harald Odland; Gábor Papp; Thomas Peitzmann; Pierluigi Piersimoni; Maksym Protsenko; Attiq Ur Rehman; Matthias Richter; Joshua Santana; Alexander Schilling; Joao Seco; Arnon Songmoolnak; Jarle Rambo Sølie; Ganesh Tambave; Ihor Tymchuk; Kjetil Ullaland; Monika Varga-Kofarago; Lennart Volz; Boris Wagner; Steffen Wendzel; Alexander Wiebel; Ren Zheng Xiao; Shiming Yang; Hiroki Yokoyama; Sebastian Zillien; Dieter Röhrich

doi:10.1080/0284186x.2021.1949037

Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks

Helge Egil Seime Pettersen^*, Max Aehle, Johan Alme, Gergely Gábor Barnaföldi, Vyacheslav Borshchov, Anthony van den Brink, Mamdouh Chaar, Viljar Eikeland, Grigory Feofilov, Christoph Garth, Nicolas R. Gauger, Georgi Genov, Ola Grøttvik, Håvard Helstrup, Sergey Igolkin, Ralf Keidel, Chinorat Kobdaj, Tobias Kortus, Viktor Leonhardt, Shruti MehendaleRaju Ningappa Mulawade, Odd Harald Odland, Gábor Papp, Thomas Peitzmann, Pierluigi Piersimoni, Maksym Protsenko, Attiq Ur Rehman, Matthias Richter, Joshua Santana, Alexander Schilling, Joao Seco, Arnon Songmoolnak, Jarle Rambo Sølie, Ganesh Tambave, Ihor Tymchuk, Kjetil Ullaland, Monika Varga-Kofarago, Lennart Volz, Boris Wagner, Steffen Wendzel, Alexander Wiebel, Ren Zheng Xiao, Shiming Yang, Hiroki Yokoyama, Sebastian Zillien, Dieter Röhrich

^*Corresponding author for this work

Subatomic Physics

Research output: Contribution to journal › Article › Academic › peer-review

Abstract

Background:
Proton computed tomography (pCT) and radiography (pRad) are proposed modalities for improved treatment plan accuracy and in situ treatment validation in proton therapy. The pCT system of the Bergen pCT collaboration is able to handle very high particle intensities by means of track reconstruction. However, incorrectly reconstructed and secondary tracks degrade the image quality. We have investigated whether a convolutional neural network (CNN)-based filter is able to improve the image quality.

Material and methods:
The CNN was trained by simulation and reconstruction of tens of millions of proton and helium tracks. The CNN filter was then compared to simple energy loss threshold methods using the Area Under the Receiver Operating Characteristics curve (AUROC), and by comparing the image quality and Water Equivalent Path Length (WEPL) error of proton and helium radiographs filtered with the same methods.

Results:
The CNN method led to a considerable improvement of the AUROC, from 74.3% to 97.5% with protons and from 94.2% to 99.5% with helium. The CNN filtering reduced the WEPL error in the helium radiograph from 1.03 mm to 0.93 mm while no improvement was seen in the CNN filtered pRads.

Conclusion:
The CNN improved the filtering of proton and helium tracks. Only in the helium radiograph did this lead to improved image quality.

Original language	English
Pages (from-to)	1413-1418
Number of pages	6
Journal	Acta Oncologica
Volume	60
Issue number	11
DOIs	https://doi.org/10.1080/0284186x.2021.1949037
Publication status	Published - 14 Jul 2021

Bibliographical note

Funding Information:
This work has been funded by the Trond Mohn Foundation [grant BFS2017TMT07] and by the Research Council of Norway [grant 250858]. We kindly acknowledge the support of the RHRK. Part of this work was supported by grants from the MWWK, Germany (research consortium SIVERT). GGB, PG, VKM are partially supported by the Hungarian Research Fund NKFIH under contracts No. K135515, 2019-2.1.11-T?T-2019-00050 and 2019-2.1.6-NEMZ_KI-2019-00011. The simulations were partly executed on the high performance cluster ?Elwetritsch? at the TU Kaiserslautern which is part of the ?Alliance of High Performance Computing Rheinland-Pfalz? (AHRP). The ALPIDE chip was developed by the ALICE collaboration at CERN.

Publisher Copyright:
© 2021 Acta Oncologica Foundation.

Keywords

convolutional neural network
machine learning
Monte Carlo simulation
Proton computed tomography
secondary particles
track reconstruction

Access to Document

10.1080/0284186x.2021.1949037

0284186X.2021Final published version, 4.26 MBLicence: Taverne

Cite this

Pettersen, H. E. S., Aehle, M., Alme, J., Barnaföldi, G. G., Borshchov, V., van den Brink, A., Chaar, M., Eikeland, V., Feofilov, G., Garth, C., Gauger, N. R., Genov, G., Grøttvik, O., Helstrup, H., Igolkin, S., Keidel, R., Kobdaj, C., Kortus, T., Leonhardt, V., ... Röhrich, D. (2021). Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks. Acta Oncologica, 60(11), 1413-1418. https://doi.org/10.1080/0284186x.2021.1949037

@article{5c3237eaf08042fc85811e5c0a0b908d,

title = "Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks",

abstract = "Background: Proton computed tomography (pCT) and radiography (pRad) are proposed modalities for improved treatment plan accuracy and in situ treatment validation in proton therapy. The pCT system of the Bergen pCT collaboration is able to handle very high particle intensities by means of track reconstruction. However, incorrectly reconstructed and secondary tracks degrade the image quality. We have investigated whether a convolutional neural network (CNN)-based filter is able to improve the image quality. Material and methods: The CNN was trained by simulation and reconstruction of tens of millions of proton and helium tracks. The CNN filter was then compared to simple energy loss threshold methods using the Area Under the Receiver Operating Characteristics curve (AUROC), and by comparing the image quality and Water Equivalent Path Length (WEPL) error of proton and helium radiographs filtered with the same methods. Results: The CNN method led to a considerable improvement of the AUROC, from 74.3% to 97.5% with protons and from 94.2% to 99.5% with helium. The CNN filtering reduced the WEPL error in the helium radiograph from 1.03 mm to 0.93 mm while no improvement was seen in the CNN filtered pRads.Conclusion: The CNN improved the filtering of proton and helium tracks. Only in the helium radiograph did this lead to improved image quality.",

keywords = "convolutional neural network, machine learning, Monte Carlo simulation, Proton computed tomography, secondary particles, track reconstruction",

author = "Pettersen, {Helge Egil Seime} and Max Aehle and Johan Alme and Barnaf{\"o}ldi, {Gergely G{\'a}bor} and Vyacheslav Borshchov and {van den Brink}, Anthony and Mamdouh Chaar and Viljar Eikeland and Grigory Feofilov and Christoph Garth and Gauger, {Nicolas R.} and Georgi Genov and Ola Gr{\o}ttvik and H{\aa}vard Helstrup and Sergey Igolkin and Ralf Keidel and Chinorat Kobdaj and Tobias Kortus and Viktor Leonhardt and Shruti Mehendale and Mulawade, {Raju Ningappa} and Odland, {Odd Harald} and G{\'a}bor Papp and Thomas Peitzmann and Pierluigi Piersimoni and Maksym Protsenko and Rehman, {Attiq Ur} and Matthias Richter and Joshua Santana and Alexander Schilling and Joao Seco and Arnon Songmoolnak and S{\o}lie, {Jarle Rambo} and Ganesh Tambave and Ihor Tymchuk and Kjetil Ullaland and Monika Varga-Kofarago and Lennart Volz and Boris Wagner and Steffen Wendzel and Alexander Wiebel and Xiao, {Ren Zheng} and Shiming Yang and Hiroki Yokoyama and Sebastian Zillien and Dieter R{\"o}hrich",

note = "Funding Information: This work has been funded by the Trond Mohn Foundation [grant BFS2017TMT07] and by the Research Council of Norway [grant 250858]. We kindly acknowledge the support of the RHRK. Part of this work was supported by grants from the MWWK, Germany (research consortium SIVERT). GGB, PG, VKM are partially supported by the Hungarian Research Fund NKFIH under contracts No. K135515, 2019-2.1.11-T?T-2019-00050 and 2019-2.1.6-NEMZ_KI-2019-00011. The simulations were partly executed on the high performance cluster ?Elwetritsch? at the TU Kaiserslautern which is part of the ?Alliance of High Performance Computing Rheinland-Pfalz? (AHRP). The ALPIDE chip was developed by the ALICE collaboration at CERN. Publisher Copyright: {\textcopyright} 2021 Acta Oncologica Foundation.",

year = "2021",

month = jul,

day = "14",

doi = "10.1080/0284186x.2021.1949037",

language = "English",

volume = "60",

pages = "1413--1418",

journal = "Acta Oncologica",

issn = "0284-186X",

publisher = "Medical Journals Sweden AB",

number = "11",

}

Pettersen, HES, Aehle, M, Alme, J, Barnaföldi, GG, Borshchov, V, van den Brink, A, Chaar, M, Eikeland, V, Feofilov, G, Garth, C, Gauger, NR, Genov, G, Grøttvik, O, Helstrup, H, Igolkin, S, Keidel, R, Kobdaj, C, Kortus, T, Leonhardt, V, Mehendale, S, Mulawade, RN, Odland, OH, Papp, G, Peitzmann, T, Piersimoni, P, Protsenko, M, Rehman, AU, Richter, M, Santana, J, Schilling, A, Seco, J, Songmoolnak, A, Sølie, JR, Tambave, G, Tymchuk, I, Ullaland, K, Varga-Kofarago, M, Volz, L, Wagner, B, Wendzel, S, Wiebel, A, Xiao, RZ, Yang, S, Yokoyama, H, Zillien, S & Röhrich, D 2021, 'Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks', Acta Oncologica, vol. 60, no. 11, pp. 1413-1418. https://doi.org/10.1080/0284186x.2021.1949037

TY - JOUR

T1 - Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks

AU - Pettersen, Helge Egil Seime

AU - Aehle, Max

AU - Alme, Johan

AU - Barnaföldi, Gergely Gábor

AU - Borshchov, Vyacheslav

AU - van den Brink, Anthony

AU - Chaar, Mamdouh

AU - Eikeland, Viljar

AU - Feofilov, Grigory

AU - Garth, Christoph

AU - Gauger, Nicolas R.

AU - Genov, Georgi

AU - Grøttvik, Ola

AU - Helstrup, Håvard

AU - Igolkin, Sergey

AU - Keidel, Ralf

AU - Kobdaj, Chinorat

AU - Kortus, Tobias

AU - Leonhardt, Viktor

AU - Mehendale, Shruti

AU - Mulawade, Raju Ningappa

AU - Odland, Odd Harald

AU - Papp, Gábor

AU - Peitzmann, Thomas

AU - Piersimoni, Pierluigi

AU - Protsenko, Maksym

AU - Rehman, Attiq Ur

AU - Richter, Matthias

AU - Santana, Joshua

AU - Schilling, Alexander

AU - Seco, Joao

AU - Songmoolnak, Arnon

AU - Sølie, Jarle Rambo

AU - Tambave, Ganesh

AU - Tymchuk, Ihor

AU - Ullaland, Kjetil

AU - Varga-Kofarago, Monika

AU - Volz, Lennart

AU - Wagner, Boris

AU - Wendzel, Steffen

AU - Wiebel, Alexander

AU - Xiao, Ren Zheng

AU - Yang, Shiming

AU - Yokoyama, Hiroki

AU - Zillien, Sebastian

AU - Röhrich, Dieter

N1 - Funding Information: This work has been funded by the Trond Mohn Foundation [grant BFS2017TMT07] and by the Research Council of Norway [grant 250858]. We kindly acknowledge the support of the RHRK. Part of this work was supported by grants from the MWWK, Germany (research consortium SIVERT). GGB, PG, VKM are partially supported by the Hungarian Research Fund NKFIH under contracts No. K135515, 2019-2.1.11-T?T-2019-00050 and 2019-2.1.6-NEMZ_KI-2019-00011. The simulations were partly executed on the high performance cluster ?Elwetritsch? at the TU Kaiserslautern which is part of the ?Alliance of High Performance Computing Rheinland-Pfalz? (AHRP). The ALPIDE chip was developed by the ALICE collaboration at CERN. Publisher Copyright: © 2021 Acta Oncologica Foundation.

PY - 2021/7/14

Y1 - 2021/7/14

N2 - Background: Proton computed tomography (pCT) and radiography (pRad) are proposed modalities for improved treatment plan accuracy and in situ treatment validation in proton therapy. The pCT system of the Bergen pCT collaboration is able to handle very high particle intensities by means of track reconstruction. However, incorrectly reconstructed and secondary tracks degrade the image quality. We have investigated whether a convolutional neural network (CNN)-based filter is able to improve the image quality. Material and methods: The CNN was trained by simulation and reconstruction of tens of millions of proton and helium tracks. The CNN filter was then compared to simple energy loss threshold methods using the Area Under the Receiver Operating Characteristics curve (AUROC), and by comparing the image quality and Water Equivalent Path Length (WEPL) error of proton and helium radiographs filtered with the same methods. Results: The CNN method led to a considerable improvement of the AUROC, from 74.3% to 97.5% with protons and from 94.2% to 99.5% with helium. The CNN filtering reduced the WEPL error in the helium radiograph from 1.03 mm to 0.93 mm while no improvement was seen in the CNN filtered pRads.Conclusion: The CNN improved the filtering of proton and helium tracks. Only in the helium radiograph did this lead to improved image quality.

AB - Background: Proton computed tomography (pCT) and radiography (pRad) are proposed modalities for improved treatment plan accuracy and in situ treatment validation in proton therapy. The pCT system of the Bergen pCT collaboration is able to handle very high particle intensities by means of track reconstruction. However, incorrectly reconstructed and secondary tracks degrade the image quality. We have investigated whether a convolutional neural network (CNN)-based filter is able to improve the image quality. Material and methods: The CNN was trained by simulation and reconstruction of tens of millions of proton and helium tracks. The CNN filter was then compared to simple energy loss threshold methods using the Area Under the Receiver Operating Characteristics curve (AUROC), and by comparing the image quality and Water Equivalent Path Length (WEPL) error of proton and helium radiographs filtered with the same methods. Results: The CNN method led to a considerable improvement of the AUROC, from 74.3% to 97.5% with protons and from 94.2% to 99.5% with helium. The CNN filtering reduced the WEPL error in the helium radiograph from 1.03 mm to 0.93 mm while no improvement was seen in the CNN filtered pRads.Conclusion: The CNN improved the filtering of proton and helium tracks. Only in the helium radiograph did this lead to improved image quality.

KW - convolutional neural network

KW - machine learning

KW - Monte Carlo simulation

KW - Proton computed tomography

KW - secondary particles

KW - track reconstruction

UR - http://www.scopus.com/inward/record.url?scp=85110794003&partnerID=8YFLogxK

U2 - 10.1080/0284186x.2021.1949037

DO - 10.1080/0284186x.2021.1949037

M3 - Article

AN - SCOPUS:85110794003

SN - 0284-186X

VL - 60

SP - 1413

EP - 1418

JO - Acta Oncologica

JF - Acta Oncologica

IS - 11

ER -

Investigating particle track topology for range telescopes in particle radiography using convolutional neural networks

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this