Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning

This study investigates the generation of synthetic CT (sCT) images from zero echo time (ZTE) MRI to support MR-only radiotherapy, which can reduce image registration errors and lower treatment planning costs. Since MRI lacks the electron density data required for accurate dose calculations, generat...

Full description

Saved in:
Bibliographic Details
Main Authors: Souha Aouadi, Mojtaba Barzegar, Alla Al-Sabahi, Tarraf Torfeh, Satheesh Paloor, Mohamed Riyas, Palmira Caparrotti, Rabih Hammoud, Noora Al-Hammadi
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Information
Subjects:
Online Access:https://www.mdpi.com/2078-2489/16/6/477
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849431849751805952
author Souha Aouadi
Mojtaba Barzegar
Alla Al-Sabahi
Tarraf Torfeh
Satheesh Paloor
Mohamed Riyas
Palmira Caparrotti
Rabih Hammoud
Noora Al-Hammadi
author_facet Souha Aouadi
Mojtaba Barzegar
Alla Al-Sabahi
Tarraf Torfeh
Satheesh Paloor
Mohamed Riyas
Palmira Caparrotti
Rabih Hammoud
Noora Al-Hammadi
author_sort Souha Aouadi
collection DOAJ
description This study investigates the generation of synthetic CT (sCT) images from zero echo time (ZTE) MRI to support MR-only radiotherapy, which can reduce image registration errors and lower treatment planning costs. Since MRI lacks the electron density data required for accurate dose calculations, generating reliable sCTs is essential. ZTE MRI, offering high bone contrast, was used with two deep learning models: attention deep residual U-Net (ADR-Unet) and derived conditional generative adversarial network (cGAN). Data from 17 head and neck cancer patients were used to train and evaluate the models. ADR-Unet was enhanced with deep residual blocks and attention mechanisms to improve learning and reconstruction quality. Both models were implemented in-house and compared to standard U-Net and Unet++ architectures using image quality metrics, visual inspection, and dosimetric analysis. Volumetric modulated arc therapy (VMAT) planning was performed on both planning CT and generated sCTs. ADR-Unet achieved a mean absolute error of 55.49 HU and a Dice score of 0.86 for bone structures. All the models demonstrated Gamma pass rates above 99.4% and dose deviations within 2–3%, confirming clinical acceptability. These results highlight ADR-Unet and cGAN as promising solutions for accurate sCT generation, enabling effective MR-only radiotherapy.
format Article
id doaj-art-b5d3ed2b1f6f46719481040dbbe3d404
institution Kabale University
issn 2078-2489
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Information
spelling doaj-art-b5d3ed2b1f6f46719481040dbbe3d4042025-08-20T03:27:29ZengMDPI AGInformation2078-24892025-06-0116647710.3390/info16060477Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep LearningSouha Aouadi0Mojtaba Barzegar1Alla Al-Sabahi2Tarraf Torfeh3Satheesh Paloor4Mohamed Riyas5Palmira Caparrotti6Rabih Hammoud7Noora Al-Hammadi8Department of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarDepartment of Radiation Oncology, National Center for Cancer Care and Research, Hamad Medical Corporation, Doha 3050, QatarThis study investigates the generation of synthetic CT (sCT) images from zero echo time (ZTE) MRI to support MR-only radiotherapy, which can reduce image registration errors and lower treatment planning costs. Since MRI lacks the electron density data required for accurate dose calculations, generating reliable sCTs is essential. ZTE MRI, offering high bone contrast, was used with two deep learning models: attention deep residual U-Net (ADR-Unet) and derived conditional generative adversarial network (cGAN). Data from 17 head and neck cancer patients were used to train and evaluate the models. ADR-Unet was enhanced with deep residual blocks and attention mechanisms to improve learning and reconstruction quality. Both models were implemented in-house and compared to standard U-Net and Unet++ architectures using image quality metrics, visual inspection, and dosimetric analysis. Volumetric modulated arc therapy (VMAT) planning was performed on both planning CT and generated sCTs. ADR-Unet achieved a mean absolute error of 55.49 HU and a Dice score of 0.86 for bone structures. All the models demonstrated Gamma pass rates above 99.4% and dose deviations within 2–3%, confirming clinical acceptability. These results highlight ADR-Unet and cGAN as promising solutions for accurate sCT generation, enabling effective MR-only radiotherapy.https://www.mdpi.com/2078-2489/16/6/477synthetic CTdeep learningZero-TE MRIMRI-only radiotherapyhead and neck
spellingShingle Souha Aouadi
Mojtaba Barzegar
Alla Al-Sabahi
Tarraf Torfeh
Satheesh Paloor
Mohamed Riyas
Palmira Caparrotti
Rabih Hammoud
Noora Al-Hammadi
Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning
Information
synthetic CT
deep learning
Zero-TE MRI
MRI-only radiotherapy
head and neck
title Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning
title_full Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning
title_fullStr Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning
title_full_unstemmed Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning
title_short Towards MR-Only Radiotherapy in Head and Neck: Generation of Synthetic CT from Zero-TE MRI Using Deep Learning
title_sort towards mr only radiotherapy in head and neck generation of synthetic ct from zero te mri using deep learning
topic synthetic CT
deep learning
Zero-TE MRI
MRI-only radiotherapy
head and neck
url https://www.mdpi.com/2078-2489/16/6/477
work_keys_str_mv AT souhaaouadi towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT mojtababarzegar towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT allaalsabahi towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT tarraftorfeh towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT satheeshpaloor towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT mohamedriyas towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT palmiracaparrotti towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT rabihhammoud towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning
AT nooraalhammadi towardsmronlyradiotherapyinheadandneckgenerationofsyntheticctfromzerotemriusingdeeplearning