Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise
<strong>Purpose</strong> To enhance the image quality of oncology [18F]-FDG PET scans acquired in shorter times and reconstructed by faster algorithms using deep neural networks. <br> <strong>Methods</strong> List-mode data from 277 [18F]-FDG PET/CT scans, from six cent...
Main Authors: | , , , , , , , , , |
---|---|
Format: | Journal article |
Language: | English |
Published: |
Springer
2021
|
_version_ | 1797050437760712704 |
---|---|
author | Mehranian, A Wollenweber, SD Walker, MD Bradley, KM Fielding, PA Su, K-H Johnsen, R Kotasidis, F Jansen, FP McGowan, D |
author_facet | Mehranian, A Wollenweber, SD Walker, MD Bradley, KM Fielding, PA Su, K-H Johnsen, R Kotasidis, F Jansen, FP McGowan, D |
author_sort | Mehranian, A |
collection | OXFORD |
description | <strong>Purpose</strong>
To enhance the image quality of oncology [18F]-FDG PET scans acquired in shorter times and reconstructed by faster algorithms using deep neural networks.
<br>
<strong>Methods</strong>
List-mode data from 277 [18F]-FDG PET/CT scans, from six centres using GE Discovery PET/CT scanners, were split into ¾-, ½- and ¼-duration scans. Full-duration datasets were reconstructed using the convergent block sequential regularised expectation maximisation (BSREM) algorithm. Short-duration datasets were reconstructed with the faster OSEM algorithm. The 277 examinations were divided into training (n = 237), validation (n = 15) and testing (n = 25) sets. Three deep learning enhancement (DLE) models were trained to map full and partial-duration OSEM images into their target full-duration BSREM images. In addition to standardised uptake value (SUV) evaluations in lesions, liver and lungs, two experienced radiologists scored the quality of testing set images and BSREM in a blinded clinical reading (175 series).
<br>
<strong>Results</strong>
OSEM reconstructions demonstrated up to 22% difference in lesion SUVmax, for different scan durations, compared to full-duration BSREM. Application of the DLE models reduced this difference significantly for full-, ¾- and ½-duration scans, while simultaneously reducing the noise in the liver. The clinical reading showed that the standard DLE model with full- or ¾-duration scans provided an image quality substantially comparable to full-duration scans with BSREM reconstruction, yet in a shorter reconstruction time.
<br>
<strong>Conclusion</strong>
Deep learning–based image enhancement models may allow a reduction in scan time (or injected activity) by up to 50%, and can decrease reconstruction time to a third, while maintaining image quality. |
first_indexed | 2024-03-06T18:05:09Z |
format | Journal article |
id | oxford-uuid:01304f26-38a5-42c8-ba1c-6ed4097014bb |
institution | University of Oxford |
language | English |
last_indexed | 2024-03-06T18:05:09Z |
publishDate | 2021 |
publisher | Springer |
record_format | dspace |
spelling | oxford-uuid:01304f26-38a5-42c8-ba1c-6ed4097014bb2022-03-26T08:33:34ZImage enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noiseJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:01304f26-38a5-42c8-ba1c-6ed4097014bbEnglishSymplectic ElementsSpringer2021Mehranian, AWollenweber, SDWalker, MDBradley, KMFielding, PASu, K-HJohnsen, RKotasidis, FJansen, FPMcGowan, D<strong>Purpose</strong> To enhance the image quality of oncology [18F]-FDG PET scans acquired in shorter times and reconstructed by faster algorithms using deep neural networks. <br> <strong>Methods</strong> List-mode data from 277 [18F]-FDG PET/CT scans, from six centres using GE Discovery PET/CT scanners, were split into ¾-, ½- and ¼-duration scans. Full-duration datasets were reconstructed using the convergent block sequential regularised expectation maximisation (BSREM) algorithm. Short-duration datasets were reconstructed with the faster OSEM algorithm. The 277 examinations were divided into training (n = 237), validation (n = 15) and testing (n = 25) sets. Three deep learning enhancement (DLE) models were trained to map full and partial-duration OSEM images into their target full-duration BSREM images. In addition to standardised uptake value (SUV) evaluations in lesions, liver and lungs, two experienced radiologists scored the quality of testing set images and BSREM in a blinded clinical reading (175 series). <br> <strong>Results</strong> OSEM reconstructions demonstrated up to 22% difference in lesion SUVmax, for different scan durations, compared to full-duration BSREM. Application of the DLE models reduced this difference significantly for full-, ¾- and ½-duration scans, while simultaneously reducing the noise in the liver. The clinical reading showed that the standard DLE model with full- or ¾-duration scans provided an image quality substantially comparable to full-duration scans with BSREM reconstruction, yet in a shorter reconstruction time. <br> <strong>Conclusion</strong> Deep learning–based image enhancement models may allow a reduction in scan time (or injected activity) by up to 50%, and can decrease reconstruction time to a third, while maintaining image quality. |
spellingShingle | Mehranian, A Wollenweber, SD Walker, MD Bradley, KM Fielding, PA Su, K-H Johnsen, R Kotasidis, F Jansen, FP McGowan, D Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise |
title | Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise |
title_full | Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise |
title_fullStr | Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise |
title_full_unstemmed | Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise |
title_short | Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise |
title_sort | image enhancement of whole body oncology 18f fdg pet scans using deep neural networks to reduce noise |
work_keys_str_mv | AT mehraniana imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT wollenwebersd imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT walkermd imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT bradleykm imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT fieldingpa imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT sukh imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT johnsenr imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT kotasidisf imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT jansenfp imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise AT mcgowand imageenhancementofwholebodyoncology18ffdgpetscansusingdeepneuralnetworkstoreducenoise |