Repository logo
 

Impact of GAN-based lesion-focused medical image super-resolution on the robustness of radiomic features

Simple item page
dc.contributor.authorde Farias, Erick Costa
dc.contributor.authordi Noia, Christian
dc.contributor.authorHan, Changhee
dc.contributor.authorSala, Evis
dc.contributor.authorCastelli, Mauro
dc.contributor.authorRundo, Leonardo
dc.date.accessioned2021-11-01T16:26:31Z
dc.date.available2021-11-01T16:26:31Z
dc.date.issued2021-11-01
dc.date.submitted2021-08-06
dc.date.updated2021-11-01T16:26:30Z
dc.description.abstractAbstract: Robust machine learning models based on radiomic features might allow for accurate diagnosis, prognosis, and medical decision-making. Unfortunately, the lack of standardized radiomic feature extraction has hampered their clinical use. Since the radiomic features tend to be affected by low voxel statistics in regions of interest, increasing the sample size would improve their robustness in clinical studies. Therefore, we propose a Generative Adversarial Network (GAN)-based lesion-focused framework for Computed Tomography (CT) image Super-Resolution (SR); for the lesion (i.e., cancer) patch-focused training, we incorporate Spatial Pyramid Pooling (SPP) into GAN-Constrained by the Identical, Residual, and Cycle Learning Ensemble (GAN-CIRCLE). At 2× SR, the proposed model achieved better perceptual quality with less blurring than the other considered state-of-the-art SR methods, while producing comparable results at 4× SR. We also evaluated the robustness of our model’s radiomic feature in terms of quantization on a different lung cancer CT dataset using Principal Component Analysis (PCA). Intriguingly, the most important radiomic features in our PCA-based analysis were the most robust features extracted on the GAN-super-resolved images. These achievements pave the way for the application of GAN-based image Super-Resolution techniques for studies of radiomics for robust biomarker discovery.
dc.identifier.doi10.17863/CAM.77582
dc.identifier.eissn2045-2322
dc.identifier.others41598-021-00898-z
dc.identifier.other898
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/330140
dc.languageen
dc.publisherNature Publishing Group UK
dc.subjectArticle
dc.subject/692/53
dc.subject/692/699/67
dc.subject/692/700/1421
dc.subject/692/308/53
dc.subject/692/4028/67/2321
dc.subject/692/4028/67/1612
dc.subject/639/166/985
dc.subjectarticle
dc.titleImpact of GAN-based lesion-focused medical image super-resolution on the robustness of radiomic features
dc.typeArticle
dcterms.dateAccepted2021-10-13
prism.issueIdentifier1
prism.publicationNameScientific Reports
prism.volume11
pubs.funder-project-idMark Foundation For Cancer Research (C9685/A25177, C9685/A25177)
pubs.funder-project-idCancer Research UK (C42780/A27066, C42780/A27066)
pubs.funder-project-idNIHR Cambridge Biomedical Research Centre (BRC-1215-20014, BRC-1215-20014)
pubs.funder-project-idWellcome Trust (215733/Z/19/Z)
pubs.funder-project-idFundação para a Ciência e a Tecnologia (DSAIPA/DS/0022/2018)
pubs.funder-project-idJavna Agencija za Raziskovalno Dejavnost RS (P5-0410)
rioxxterms.licenseref.urihttp://creativecommons.org/licenses/by/4.0/
rioxxterms.versionVoR
rioxxterms.versionofrecord10.1038/s41598-021-00898-z

Files

Original bundle
Now showing 1 - 2 of 2
No Thumbnail Available
Name:
41598_2021_Article_898_nlm.xml
Size:
156.51 KB
Format:
Extensible Markup Language
Description:
Bibliographic metadata
Licence
http://creativecommons.org/licenses/by/4.0/
Download
Thumbnail Image
Name:
41598_2021_Article_898.pdf
Size:
2.56 MB
Format:
Adobe Portable Document Format
Description:
Published version
Licence
http://creativecommons.org/licenses/by/4.0/
Download

Collections

Jisc Publications Router