Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report

Evi M.C. Huijben; Maarten L. Terpstra; Arthur Jr. Galapon; Suraj Pai; Adrian Thummerer; Peter Koopmans; Manya Afonso; Maureen van Eijnatten; Oliver Gurney-Champion; Zeli Chen; Yiwen Zhang; Kaiyi Zheng; Chuanpu Li; Haowen Pang; Chuyang Ye; Runqi Wang; Tao Song; Fuxin Fan; Jingna Qiu; Yixing Huang; Juhyung Ha; Jong Sung Park; Alexandra Alain-Beaudoin; Silvain Bériault; Pengxin Yu; Hongbin Guo; Zhanyao Huang; Gengwan Li; Xueru Zhang; Yubo Fan; Han Liu; Bowen Xin; Aaron Nicolson; Lujia Zhong; Zhiwei Deng; Gustav Müller-Franzes; Firas Khader; Xia Li; Ye Zhang; Cédric Hémon; Valentin Boussot; Zhihao Zhang; Long Wang; Lu Bai; Shaobin Wang; Derk Mus; Bram Kooiman; Chelsea A.H. Sargeant; Edward G.A. Henderson; Satoshi Kondo; Satoshi Kasai; Reza Karimzadeh; Bulat Ibragimov; Thomas Helfer; Jessica Dafflon; Zijie Chen; Enpei Wang; Zoltan Perko; Matteo Maspero

doi:10.1016/j.media.2024.103276

Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report

Evi M.C. Huijben
, Maarten L. Terpstra
, Arthur Jr. Galapon
, Suraj Pai
, Adrian Thummerer
, Peter Koopmans
, Manya Afonso
, Maureen van Eijnatten
, Oliver Gurney-Champion
, Zeli Chen
, Yiwen Zhang
, Kaiyi Zheng
, Chuanpu Li
, Haowen Pang
, Chuyang Ye
, Runqi Wang
, Tao Song
, Fuxin Fan
, Jingna Qiu
, Yixing Huang

Juhyung Ha, Jong Sung Park, Alexandra Alain-Beaudoin, Silvain Bériault, Pengxin Yu, Hongbin Guo, Zhanyao Huang, Gengwan Li, Xueru Zhang, Yubo Fan, Han Liu, Bowen Xin, Aaron Nicolson, Lujia Zhong, Zhiwei Deng, Gustav Müller-Franzes, Firas Khader, Xia Li, Ye Zhang, Cédric Hémon, Valentin Boussot, Zhihao Zhang, Long Wang, Lu Bai, Shaobin Wang, Derk Mus, Bram Kooiman, Chelsea A.H. Sargeant, Edward G.A. Henderson, Satoshi Kondo, Satoshi Kasai, Reza Karimzadeh, Bulat Ibragimov, Thomas Helfer, Jessica Dafflon, Zijie Chen, Enpei Wang, Zoltan Perko, Matteo Maspero (Corresponding author)

Research output: Contribution to journal › Short survey › peer-review

29 Citations (Scopus)

27 Downloads (Pure)

Abstract

Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information, while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: (1) MRI-to-CT and (2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (≥0.87/0.90) and gamma pass rates for photon (≥98.1%/99.0%) and proton (≥97.3%/97.0%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy. It showcased the growing capacity of deep learning to produce high-quality sCT, reducing reliance on conventional CT for treatment planning.

Original language	English
Article number	103276
Number of pages	21
Journal	Medical Image Analysis
Volume	97
DOIs	https://doi.org/10.1016/j.media.2024.103276
Publication status	Published - Oct 2024

Bibliographical note

Publisher Copyright:
© 2024 The Author(s)

Keywords

Deep learning
Medical image synthesis
Radiotherapy
Synthetic CT generation
Radiotherapy Dosage
Neoplasms/radiotherapy
Radiotherapy Planning, Computer-Assisted/methods
Radiotherapy, Image-Guided/methods
Tomography, X-Ray Computed/methods
Humans
Magnetic Resonance Imaging/methods
Cone-Beam Computed Tomography/methods

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.media.2024.103276

pdfFinal published version, 2.18 MBLicence: CC BY

Cite this

Huijben, E. M. C., Terpstra, M. L., Galapon, A. J., Pai, S., Thummerer, A., Koopmans, P., Afonso, M., van Eijnatten, M., Gurney-Champion, O., Chen, Z., Zhang, Y., Zheng, K., Li, C., Pang, H., Ye, C., Wang, R., Song, T., Fan, F., Qiu, J., ... Maspero, M. (2024). Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report. Medical Image Analysis, 97, Article 103276. https://doi.org/10.1016/j.media.2024.103276

@article{51a19e744b3d4969ac27bb0e82af10c1,

title = "Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report",

abstract = "Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information, while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: (1) MRI-to-CT and (2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (≥0.87/0.90) and gamma pass rates for photon (≥98.1\%/99.0\%) and proton (≥97.3\%/97.0\%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy. It showcased the growing capacity of deep learning to produce high-quality sCT, reducing reliance on conventional CT for treatment planning.",

keywords = "Deep learning, Medical image synthesis, Radiotherapy, Synthetic CT generation, Radiotherapy Dosage, Neoplasms/radiotherapy, Radiotherapy Planning, Computer-Assisted/methods, Radiotherapy, Image-Guided/methods, Tomography, X-Ray Computed/methods, Humans, Magnetic Resonance Imaging/methods, Cone-Beam Computed Tomography/methods",

author = "Huijben, \{Evi M.C.\} and Terpstra, \{Maarten L.\} and Galapon, \{Arthur Jr.\} and Suraj Pai and Adrian Thummerer and Peter Koopmans and Manya Afonso and \{van Eijnatten\}, Maureen and Oliver Gurney-Champion and Zeli Chen and Yiwen Zhang and Kaiyi Zheng and Chuanpu Li and Haowen Pang and Chuyang Ye and Runqi Wang and Tao Song and Fuxin Fan and Jingna Qiu and Yixing Huang and Juhyung Ha and \{Sung Park\}, Jong and Alexandra Alain-Beaudoin and Silvain B{\'e}riault and Pengxin Yu and Hongbin Guo and Zhanyao Huang and Gengwan Li and Xueru Zhang and Yubo Fan and Han Liu and Bowen Xin and Aaron Nicolson and Lujia Zhong and Zhiwei Deng and Gustav M{\"u}ller-Franzes and Firas Khader and Xia Li and Ye Zhang and C{\'e}dric H{\'e}mon and Valentin Boussot and Zhihao Zhang and Long Wang and Lu Bai and Shaobin Wang and Derk Mus and Bram Kooiman and Sargeant, \{Chelsea A.H.\} and Henderson, \{Edward G.A.\} and Satoshi Kondo and Satoshi Kasai and Reza Karimzadeh and Bulat Ibragimov and Thomas Helfer and Jessica Dafflon and Zijie Chen and Enpei Wang and Zoltan Perko and Matteo Maspero",

note = "Publisher Copyright: {\textcopyright} 2024 The Author(s)",

year = "2024",

month = oct,

doi = "10.1016/j.media.2024.103276",

language = "English",

volume = "97",

journal = "Medical Image Analysis",

issn = "1361-8415",

publisher = "Elsevier B.V.",

}

Huijben, EMC, Terpstra, ML, Galapon, AJ, Pai, S, Thummerer, A, Koopmans, P, Afonso, M, van Eijnatten, M, Gurney-Champion, O, Chen, Z, Zhang, Y, Zheng, K, Li, C, Pang, H, Ye, C, Wang, R, Song, T, Fan, F, Qiu, J, Huang, Y, Ha, J, Sung Park, J, Alain-Beaudoin, A, Bériault, S, Yu, P, Guo, H, Huang, Z, Li, G, Zhang, X, Fan, Y, Liu, H, Xin, B, Nicolson, A, Zhong, L, Deng, Z, Müller-Franzes, G, Khader, F, Li, X, Zhang, Y, Hémon, C, Boussot, V, Zhang, Z, Wang, L, Bai, L, Wang, S, Mus, D, Kooiman, B, Sargeant, CAH, Henderson, EGA, Kondo, S, Kasai, S, Karimzadeh, R, Ibragimov, B, Helfer, T, Dafflon, J, Chen, Z, Wang, E, Perko, Z & Maspero, M 2024, 'Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report', Medical Image Analysis, vol. 97, 103276. https://doi.org/10.1016/j.media.2024.103276

TY - JOUR

T1 - Generating synthetic computed tomography for radiotherapy

T2 - SynthRAD2023 challenge report

AU - Huijben, Evi M.C.

AU - Terpstra, Maarten L.

AU - Galapon, Arthur Jr.

AU - Pai, Suraj

AU - Thummerer, Adrian

AU - Koopmans, Peter

AU - Afonso, Manya

AU - van Eijnatten, Maureen

AU - Gurney-Champion, Oliver

AU - Chen, Zeli

AU - Zhang, Yiwen

AU - Zheng, Kaiyi

AU - Li, Chuanpu

AU - Pang, Haowen

AU - Ye, Chuyang

AU - Wang, Runqi

AU - Song, Tao

AU - Fan, Fuxin

AU - Qiu, Jingna

AU - Huang, Yixing

AU - Ha, Juhyung

AU - Sung Park, Jong

AU - Alain-Beaudoin, Alexandra

AU - Bériault, Silvain

AU - Yu, Pengxin

AU - Guo, Hongbin

AU - Huang, Zhanyao

AU - Li, Gengwan

AU - Zhang, Xueru

AU - Fan, Yubo

AU - Liu, Han

AU - Xin, Bowen

AU - Nicolson, Aaron

AU - Zhong, Lujia

AU - Deng, Zhiwei

AU - Müller-Franzes, Gustav

AU - Khader, Firas

AU - Li, Xia

AU - Zhang, Ye

AU - Hémon, Cédric

AU - Boussot, Valentin

AU - Zhang, Zhihao

AU - Wang, Long

AU - Bai, Lu

AU - Wang, Shaobin

AU - Mus, Derk

AU - Kooiman, Bram

AU - Sargeant, Chelsea A.H.

AU - Henderson, Edward G.A.

AU - Kondo, Satoshi

AU - Kasai, Satoshi

AU - Karimzadeh, Reza

AU - Ibragimov, Bulat

AU - Helfer, Thomas

AU - Dafflon, Jessica

AU - Chen, Zijie

AU - Wang, Enpei

AU - Perko, Zoltan

AU - Maspero, Matteo

PY - 2024/10

Y1 - 2024/10

N2 - Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information, while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: (1) MRI-to-CT and (2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (≥0.87/0.90) and gamma pass rates for photon (≥98.1%/99.0%) and proton (≥97.3%/97.0%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy. It showcased the growing capacity of deep learning to produce high-quality sCT, reducing reliance on conventional CT for treatment planning.

AB - Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information, while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: (1) MRI-to-CT and (2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (≥0.87/0.90) and gamma pass rates for photon (≥98.1%/99.0%) and proton (≥97.3%/97.0%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy. It showcased the growing capacity of deep learning to produce high-quality sCT, reducing reliance on conventional CT for treatment planning.

KW - Deep learning

KW - Medical image synthesis

KW - Radiotherapy

KW - Synthetic CT generation

KW - Radiotherapy Dosage

KW - Neoplasms/radiotherapy

KW - Radiotherapy Planning, Computer-Assisted/methods

KW - Radiotherapy, Image-Guided/methods

KW - Tomography, X-Ray Computed/methods

KW - Humans

KW - Magnetic Resonance Imaging/methods

KW - Cone-Beam Computed Tomography/methods

UR - https://www.scopus.com/pages/publications/85199571147

U2 - 10.1016/j.media.2024.103276

DO - 10.1016/j.media.2024.103276

M3 - Short survey

C2 - 39068830

AN - SCOPUS:85199571147

SN - 1361-8415

VL - 97

JO - Medical Image Analysis

JF - Medical Image Analysis

M1 - 103276

ER -

Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report

Abstract

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this