Iterative framework for the joint segmentation and CT synthesis of MR images: Application to MRI-only radiotherapy treatment planning

Ninon Burgos; Filipa Guerreiro; Jamie McClelland; Benoit Presles; Marc Modat; Simeon Nill; David Dearnaley; Nandita DeSouza; Uwe Oelfke; Antje Christin Knopf; Sebastien Ourselin; M. Jorge Cardoso

doi:10.1088/1361-6560/aa66bf

Iterative framework for the joint segmentation and CT synthesis of MR images: Application to MRI-only radiotherapy treatment planning

Ninon Burgos, Filipa Guerreiro, Jamie McClelland, Benoit Presles, Marc Modat, Simeon Nill, David Dearnaley, Nandita DeSouza, Uwe Oelfke, Antje Christin Knopf, Sebastien Ourselin, M. Jorge Cardoso

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

38 Citations (Scopus)

79 Downloads (Pure)

Abstract

To tackle the problem of magnetic resonance imaging (MRI)-only radiotherapy treatment planning (RTP), we propose a multi-atlas information propagation scheme that jointly segments organs and generates pseudo x-ray computed tomography (CT) data from structural MR images (T1-weighted and T2-weighted). As the performance of the method strongly depends on the quality of the atlas database composed of multiple sets of aligned MR, CT and segmented images, we also propose a robust way of registering atlas MR and CT images, which combines structure-guided registration, and CT and MR image synthesis. We first evaluated the proposed framework in terms of segmentation and CT synthesis accuracy on 15 subjects with prostate cancer. The segmentations obtained with the proposed method were compared using the Dice score coefficient (DSC) to the manual segmentations. Mean DSCs of 0.73, 0.90, 0.77 and 0.90 were obtained for the prostate, bladder, rectum and femur heads, respectively. The mean absolute error (MAE) and the mean error (ME) were computed between the reference CTs (non-rigidly aligned to the MRs) and the pseudo CTs generated with the proposed method. The MAE was on average 45.7±4.6 HU and the ME -1.6±7.7 HU. We then performed a dosimetric evaluation by re-calculating plans on the pseudo CTs and comparing them to the plans optimised on the reference CTs. We compared the cumulative dose volume histograms (DVH) obtained for the pseudo CTs to the DVH obtained for the reference CTs in the planning target volume (PTV) located in the prostate, and in the organs at risk at different DVH points. We obtained average differences of -0.14% in the PTV for D_98%, and between -0.14% and 0.05% in the PTV, bladder, rectum and femur heads for D_mean and D_2%. Overall, we demonstrate that the proposed framework is able to automatically generate accurate pseudo CT images and segmentations in the pelvic region, potentially bypassing the need for CT scan for accurate RTP.

Original language	English
Pages (from-to)	4237-4253
Number of pages	17
Journal	Physics in Medicine and Biology
Volume	62
Issue number	11
Early online date	5 May 2017
DOIs	https://doi.org/10.1088/1361-6560/aa66bf
Publication status	Published - 5 May 2017

Keywords

atlas-based methods
image synthesis
MRI-only RTP
pseudo CT
segmentation

UN SDGs

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

Access to Document

10.1088/1361-6560/aa66bfLicence: CC BY

Iterative framework for the_BURGOS_5May2017_GOLD VoR (CC BY)Final published version, 3.22 MBLicence: CC BY

Cite this

Burgos, N., Guerreiro, F., McClelland, J., Presles, B., Modat, M., Nill, S., Dearnaley, D., DeSouza, N., Oelfke, U., Knopf, A. C., Ourselin, S., & Jorge Cardoso, M. (2017). Iterative framework for the joint segmentation and CT synthesis of MR images: Application to MRI-only radiotherapy treatment planning. Physics in Medicine and Biology, 62(11), 4237-4253. https://doi.org/10.1088/1361-6560/aa66bf

@article{475dfd2b6c6e449097ee84bde59e2262,

title = "Iterative framework for the joint segmentation and CT synthesis of MR images: Application to MRI-only radiotherapy treatment planning",

abstract = "To tackle the problem of magnetic resonance imaging (MRI)-only radiotherapy treatment planning (RTP), we propose a multi-atlas information propagation scheme that jointly segments organs and generates pseudo x-ray computed tomography (CT) data from structural MR images (T1-weighted and T2-weighted). As the performance of the method strongly depends on the quality of the atlas database composed of multiple sets of aligned MR, CT and segmented images, we also propose a robust way of registering atlas MR and CT images, which combines structure-guided registration, and CT and MR image synthesis. We first evaluated the proposed framework in terms of segmentation and CT synthesis accuracy on 15 subjects with prostate cancer. The segmentations obtained with the proposed method were compared using the Dice score coefficient (DSC) to the manual segmentations. Mean DSCs of 0.73, 0.90, 0.77 and 0.90 were obtained for the prostate, bladder, rectum and femur heads, respectively. The mean absolute error (MAE) and the mean error (ME) were computed between the reference CTs (non-rigidly aligned to the MRs) and the pseudo CTs generated with the proposed method. The MAE was on average 45.7±4.6 HU and the ME -1.6±7.7 HU. We then performed a dosimetric evaluation by re-calculating plans on the pseudo CTs and comparing them to the plans optimised on the reference CTs. We compared the cumulative dose volume histograms (DVH) obtained for the pseudo CTs to the DVH obtained for the reference CTs in the planning target volume (PTV) located in the prostate, and in the organs at risk at different DVH points. We obtained average differences of -0.14% in the PTV for D98%, and between -0.14% and 0.05% in the PTV, bladder, rectum and femur heads for Dmean and D2%. Overall, we demonstrate that the proposed framework is able to automatically generate accurate pseudo CT images and segmentations in the pelvic region, potentially bypassing the need for CT scan for accurate RTP.",

keywords = "atlas-based methods, image synthesis, MRI-only RTP, pseudo CT, segmentation",

author = "Ninon Burgos and Filipa Guerreiro and Jamie McClelland and Benoit Presles and Marc Modat and Simeon Nill and David Dearnaley and Nandita DeSouza and Uwe Oelfke and Knopf, {Antje Christin} and Sebastien Ourselin and {Jorge Cardoso}, M.",

year = "2017",

month = may,

day = "5",

doi = "10.1088/1361-6560/aa66bf",

language = "English",

volume = "62",

pages = "4237--4253",

journal = "Physics in Medicine and Biology",

issn = "0031-9155",

publisher = "IOP Publishing Ltd.",

number = "11",

}

Burgos, N, Guerreiro, F, McClelland, J, Presles, B, Modat, M, Nill, S, Dearnaley, D, DeSouza, N, Oelfke, U, Knopf, AC, Ourselin, S & Jorge Cardoso, M 2017, 'Iterative framework for the joint segmentation and CT synthesis of MR images: Application to MRI-only radiotherapy treatment planning', Physics in Medicine and Biology, vol. 62, no. 11, pp. 4237-4253. https://doi.org/10.1088/1361-6560/aa66bf

TY - JOUR

T1 - Iterative framework for the joint segmentation and CT synthesis of MR images

T2 - Application to MRI-only radiotherapy treatment planning

AU - Burgos, Ninon

AU - Guerreiro, Filipa

AU - McClelland, Jamie

AU - Presles, Benoit

AU - Modat, Marc

AU - Nill, Simeon

AU - Dearnaley, David

AU - DeSouza, Nandita

AU - Oelfke, Uwe

AU - Knopf, Antje Christin

AU - Ourselin, Sebastien

AU - Jorge Cardoso, M.

PY - 2017/5/5

Y1 - 2017/5/5

N2 - To tackle the problem of magnetic resonance imaging (MRI)-only radiotherapy treatment planning (RTP), we propose a multi-atlas information propagation scheme that jointly segments organs and generates pseudo x-ray computed tomography (CT) data from structural MR images (T1-weighted and T2-weighted). As the performance of the method strongly depends on the quality of the atlas database composed of multiple sets of aligned MR, CT and segmented images, we also propose a robust way of registering atlas MR and CT images, which combines structure-guided registration, and CT and MR image synthesis. We first evaluated the proposed framework in terms of segmentation and CT synthesis accuracy on 15 subjects with prostate cancer. The segmentations obtained with the proposed method were compared using the Dice score coefficient (DSC) to the manual segmentations. Mean DSCs of 0.73, 0.90, 0.77 and 0.90 were obtained for the prostate, bladder, rectum and femur heads, respectively. The mean absolute error (MAE) and the mean error (ME) were computed between the reference CTs (non-rigidly aligned to the MRs) and the pseudo CTs generated with the proposed method. The MAE was on average 45.7±4.6 HU and the ME -1.6±7.7 HU. We then performed a dosimetric evaluation by re-calculating plans on the pseudo CTs and comparing them to the plans optimised on the reference CTs. We compared the cumulative dose volume histograms (DVH) obtained for the pseudo CTs to the DVH obtained for the reference CTs in the planning target volume (PTV) located in the prostate, and in the organs at risk at different DVH points. We obtained average differences of -0.14% in the PTV for D98%, and between -0.14% and 0.05% in the PTV, bladder, rectum and femur heads for Dmean and D2%. Overall, we demonstrate that the proposed framework is able to automatically generate accurate pseudo CT images and segmentations in the pelvic region, potentially bypassing the need for CT scan for accurate RTP.

AB - To tackle the problem of magnetic resonance imaging (MRI)-only radiotherapy treatment planning (RTP), we propose a multi-atlas information propagation scheme that jointly segments organs and generates pseudo x-ray computed tomography (CT) data from structural MR images (T1-weighted and T2-weighted). As the performance of the method strongly depends on the quality of the atlas database composed of multiple sets of aligned MR, CT and segmented images, we also propose a robust way of registering atlas MR and CT images, which combines structure-guided registration, and CT and MR image synthesis. We first evaluated the proposed framework in terms of segmentation and CT synthesis accuracy on 15 subjects with prostate cancer. The segmentations obtained with the proposed method were compared using the Dice score coefficient (DSC) to the manual segmentations. Mean DSCs of 0.73, 0.90, 0.77 and 0.90 were obtained for the prostate, bladder, rectum and femur heads, respectively. The mean absolute error (MAE) and the mean error (ME) were computed between the reference CTs (non-rigidly aligned to the MRs) and the pseudo CTs generated with the proposed method. The MAE was on average 45.7±4.6 HU and the ME -1.6±7.7 HU. We then performed a dosimetric evaluation by re-calculating plans on the pseudo CTs and comparing them to the plans optimised on the reference CTs. We compared the cumulative dose volume histograms (DVH) obtained for the pseudo CTs to the DVH obtained for the reference CTs in the planning target volume (PTV) located in the prostate, and in the organs at risk at different DVH points. We obtained average differences of -0.14% in the PTV for D98%, and between -0.14% and 0.05% in the PTV, bladder, rectum and femur heads for Dmean and D2%. Overall, we demonstrate that the proposed framework is able to automatically generate accurate pseudo CT images and segmentations in the pelvic region, potentially bypassing the need for CT scan for accurate RTP.

KW - atlas-based methods

KW - image synthesis

KW - MRI-only RTP

KW - pseudo CT

KW - segmentation

UR - http://www.scopus.com/inward/record.url?scp=85020026530&partnerID=8YFLogxK

U2 - 10.1088/1361-6560/aa66bf

DO - 10.1088/1361-6560/aa66bf

M3 - Article

C2 - 28291745

AN - SCOPUS:85020026530

SN - 0031-9155

VL - 62

SP - 4237

EP - 4253

JO - Physics in Medicine and Biology

JF - Physics in Medicine and Biology

IS - 11

ER -

Iterative framework for the joint segmentation and CT synthesis of MR images: Application to MRI-only radiotherapy treatment planning

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this