MRI-only brain radiotherapy: Assessing the dosimetric accuracy of synthetic CT images generated using a deep learning approach

Samaneh Kazemifar, Sarah McGuire, Robert Timmerman, Zabi Wardak, Dan Nguyen, Yang Park, Steve Jiang, Amir Owrangi

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Purpose: This study assessed the dosimetric accuracy of synthetic CT images generated from magnetic resonance imaging (MRI) data for focal brain radiation therapy, using a deep learning approach. Material and methods: We conducted a study in 77 patients with brain tumors who had undergone both MRI and computed tomography (CT) imaging as part of their simulation for external beam treatment planning. We designed a generative adversarial network (GAN) to generate synthetic CT images from MRI images. We used Mutual Information (MI) as the loss function in the generator to overcome the misalignment between MRI and CT images (unregistered data). The model was trained using all MRI slices with corresponding CT slices from each training subject's MRI/CT pair. Results: The proposed GAN method produced an average mean absolute error (MAE) of 47.2 ± 11.0 HU over 5-fold cross validation. The overall mean Dice similarity coefficient between CT and synthetic CT images was 80% ± 6% in bone for all test data. Though training a GAN model may take several hours, the model only needs to be trained once. Generating a complete synthetic CT volume for each new patient MRI volume using a trained GAN model took only one second. Conclusions: The GAN model we developed produced highly accurate synthetic CT images from conventional, single-sequence MRI images in seconds. Our proposed method has strong potential to perform well in a clinical workflow for MRI-only brain treatment planning.

Original languageEnglish (US)
Pages (from-to)56-63
Number of pages8
JournalRadiotherapy and Oncology
Volume136
DOIs
StatePublished - Jul 1 2019

Fingerprint

Radiotherapy
Tomography
Magnetic Resonance Imaging
Learning
Brain
Cone-Beam Computed Tomography
Workflow
Brain Neoplasms
Bone and Bones
Therapeutics

Keywords

  • Brain MRI
  • Deep learning
  • GAN structure
  • Mutual information
  • Synthetic CT

ASJC Scopus subject areas

  • Hematology
  • Oncology
  • Radiology Nuclear Medicine and imaging

Cite this

@article{0f52e6e34bcf49bebf31430316e29a1a,
title = "MRI-only brain radiotherapy: Assessing the dosimetric accuracy of synthetic CT images generated using a deep learning approach",
abstract = "Purpose: This study assessed the dosimetric accuracy of synthetic CT images generated from magnetic resonance imaging (MRI) data for focal brain radiation therapy, using a deep learning approach. Material and methods: We conducted a study in 77 patients with brain tumors who had undergone both MRI and computed tomography (CT) imaging as part of their simulation for external beam treatment planning. We designed a generative adversarial network (GAN) to generate synthetic CT images from MRI images. We used Mutual Information (MI) as the loss function in the generator to overcome the misalignment between MRI and CT images (unregistered data). The model was trained using all MRI slices with corresponding CT slices from each training subject's MRI/CT pair. Results: The proposed GAN method produced an average mean absolute error (MAE) of 47.2 ± 11.0 HU over 5-fold cross validation. The overall mean Dice similarity coefficient between CT and synthetic CT images was 80{\%} ± 6{\%} in bone for all test data. Though training a GAN model may take several hours, the model only needs to be trained once. Generating a complete synthetic CT volume for each new patient MRI volume using a trained GAN model took only one second. Conclusions: The GAN model we developed produced highly accurate synthetic CT images from conventional, single-sequence MRI images in seconds. Our proposed method has strong potential to perform well in a clinical workflow for MRI-only brain treatment planning.",
keywords = "Brain MRI, Deep learning, GAN structure, Mutual information, Synthetic CT",
author = "Samaneh Kazemifar and Sarah McGuire and Robert Timmerman and Zabi Wardak and Dan Nguyen and Yang Park and Steve Jiang and Amir Owrangi",
year = "2019",
month = "7",
day = "1",
doi = "10.1016/j.radonc.2019.03.026",
language = "English (US)",
volume = "136",
pages = "56--63",
journal = "Radiotherapy and Oncology",
issn = "0167-8140",
publisher = "Elsevier Ireland Ltd",

}

TY - JOUR

T1 - MRI-only brain radiotherapy

T2 - Assessing the dosimetric accuracy of synthetic CT images generated using a deep learning approach

AU - Kazemifar, Samaneh

AU - McGuire, Sarah

AU - Timmerman, Robert

AU - Wardak, Zabi

AU - Nguyen, Dan

AU - Park, Yang

AU - Jiang, Steve

AU - Owrangi, Amir

PY - 2019/7/1

Y1 - 2019/7/1

N2 - Purpose: This study assessed the dosimetric accuracy of synthetic CT images generated from magnetic resonance imaging (MRI) data for focal brain radiation therapy, using a deep learning approach. Material and methods: We conducted a study in 77 patients with brain tumors who had undergone both MRI and computed tomography (CT) imaging as part of their simulation for external beam treatment planning. We designed a generative adversarial network (GAN) to generate synthetic CT images from MRI images. We used Mutual Information (MI) as the loss function in the generator to overcome the misalignment between MRI and CT images (unregistered data). The model was trained using all MRI slices with corresponding CT slices from each training subject's MRI/CT pair. Results: The proposed GAN method produced an average mean absolute error (MAE) of 47.2 ± 11.0 HU over 5-fold cross validation. The overall mean Dice similarity coefficient between CT and synthetic CT images was 80% ± 6% in bone for all test data. Though training a GAN model may take several hours, the model only needs to be trained once. Generating a complete synthetic CT volume for each new patient MRI volume using a trained GAN model took only one second. Conclusions: The GAN model we developed produced highly accurate synthetic CT images from conventional, single-sequence MRI images in seconds. Our proposed method has strong potential to perform well in a clinical workflow for MRI-only brain treatment planning.

AB - Purpose: This study assessed the dosimetric accuracy of synthetic CT images generated from magnetic resonance imaging (MRI) data for focal brain radiation therapy, using a deep learning approach. Material and methods: We conducted a study in 77 patients with brain tumors who had undergone both MRI and computed tomography (CT) imaging as part of their simulation for external beam treatment planning. We designed a generative adversarial network (GAN) to generate synthetic CT images from MRI images. We used Mutual Information (MI) as the loss function in the generator to overcome the misalignment between MRI and CT images (unregistered data). The model was trained using all MRI slices with corresponding CT slices from each training subject's MRI/CT pair. Results: The proposed GAN method produced an average mean absolute error (MAE) of 47.2 ± 11.0 HU over 5-fold cross validation. The overall mean Dice similarity coefficient between CT and synthetic CT images was 80% ± 6% in bone for all test data. Though training a GAN model may take several hours, the model only needs to be trained once. Generating a complete synthetic CT volume for each new patient MRI volume using a trained GAN model took only one second. Conclusions: The GAN model we developed produced highly accurate synthetic CT images from conventional, single-sequence MRI images in seconds. Our proposed method has strong potential to perform well in a clinical workflow for MRI-only brain treatment planning.

KW - Brain MRI

KW - Deep learning

KW - GAN structure

KW - Mutual information

KW - Synthetic CT

UR - http://www.scopus.com/inward/record.url?scp=85064275699&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064275699&partnerID=8YFLogxK

U2 - 10.1016/j.radonc.2019.03.026

DO - 10.1016/j.radonc.2019.03.026

M3 - Article

C2 - 31015130

AN - SCOPUS:85064275699

VL - 136

SP - 56

EP - 63

JO - Radiotherapy and Oncology

JF - Radiotherapy and Oncology

SN - 0167-8140

ER -