File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
  • Find via Find It@HKUL

Article: Pushing the limits of low-cost ultralow-field MRI by dual-acquisition deep learning 3D superresolution

TitlePushing the limits of low-cost ultralow-field MRI by dual-acquisition deep learning 3D superresolution
Authors
Issue Date1-Apr-2023
PublisherWiley
Citation
Magnetic Resonance in Medicine, 2023 How to Cite?
Abstract

Purpose

Recent development of ultralow-field (ULF) MRI presents opportunities for low-power, shielding-free, and portable clinical applications at a fraction of the cost. However, its performance remains limited by poor image quality. Here, a computational approach is formulated to advance ULF MR brain imaging through deep learning of large-scale publicly available 3T brain data.

Methods

A dual-acquisition 3D superresolution model is developed for ULF brain MRI at 0.055 T. It consists of deep cross-scale feature extraction, attentional fusion of two acquisitions, and reconstruction. Models for T1-weighted and T2-weighted imaging were trained with 3D ULF image data sets synthesized from the high-resolution 3T brain data from the Human Connectome Project. They were applied to 0.055T brain MRI with two repetitions and isotropic 3-mm acquisition resolution in healthy volunteers, young and old, as well as patients.

Results

The proposed approach significantly enhanced image spatial resolution and suppressed noise/artifacts. It yielded high 3D image quality at 0.055 T for the two most common neuroimaging protocols with isotropic 1.5-mm synthetic resolution and total scan time under 20 min. Fine anatomical details were restored with intrasubject reproducibility, intercontrast consistency, and confirmed by 3T MRI.

Conclusion

The proposed dual-acquisition 3D supe-resolution approach advances ULF MRI for quality brain imaging through deep learning of high-field brain data. Such strategy can empower ULF MRI for low-cost brain imaging, especially in point-of-care scenarios or/and in low-income and mid-income countries.


Persistent Identifierhttp://hdl.handle.net/10722/328305
ISSN
2023 Impact Factor: 3.0
2023 SCImago Journal Rankings: 1.343

 

DC FieldValueLanguage
dc.contributor.authorLau, V-
dc.contributor.authorXiao, LF-
dc.contributor.authorZhao, YJ-
dc.contributor.authorSu, S-
dc.contributor.authorDing, Y-
dc.contributor.authorMan, C-
dc.contributor.authorWang, XD-
dc.contributor.authorTsang, A-
dc.contributor.authorCao, P-
dc.contributor.authorLau, GKK-
dc.contributor.authorLeung, GKK-
dc.contributor.authorLeong, ATL-
dc.contributor.authorWu, EX-
dc.date.accessioned2023-06-28T04:41:45Z-
dc.date.available2023-06-28T04:41:45Z-
dc.date.issued2023-04-01-
dc.identifier.citationMagnetic Resonance in Medicine, 2023-
dc.identifier.issn0740-3194-
dc.identifier.urihttp://hdl.handle.net/10722/328305-
dc.description.abstract<h3>Purpose</h3><p>Recent development of ultralow-field (ULF) MRI presents opportunities for low-power, shielding-free, and portable clinical applications at a fraction of the cost. However, its performance remains limited by poor image quality. Here, a computational approach is formulated to advance ULF MR brain imaging through deep learning of large-scale publicly available 3T brain data.</p><h3>Methods</h3><p>A dual-acquisition 3D superresolution model is developed for ULF brain MRI at 0.055 T. It consists of deep cross-scale feature extraction, attentional fusion of two acquisitions, and reconstruction. Models for T<sub>1</sub>-weighted and T<sub>2</sub>-weighted imaging were trained with 3D ULF image data sets synthesized from the high-resolution 3T brain data from the Human Connectome Project. They were applied to 0.055T brain MRI with two repetitions and isotropic 3-mm acquisition resolution in healthy volunteers, young and old, as well as patients.</p><h3>Results</h3><p>The proposed approach significantly enhanced image spatial resolution and suppressed noise/artifacts. It yielded high 3D image quality at 0.055 T for the two most common neuroimaging protocols with isotropic 1.5-mm synthetic resolution and total scan time under 20 min. Fine anatomical details were restored with intrasubject reproducibility, intercontrast consistency, and confirmed by 3T MRI.</p><h3>Conclusion</h3><p>The proposed dual-acquisition 3D supe-resolution approach advances ULF MRI for quality brain imaging through deep learning of high-field brain data. Such strategy can empower ULF MRI for low-cost brain imaging, especially in point-of-care scenarios or/and in low-income and mid-income countries.</p>-
dc.languageeng-
dc.publisherWiley-
dc.relation.ispartofMagnetic Resonance in Medicine-
dc.titlePushing the limits of low-cost ultralow-field MRI by dual-acquisition deep learning 3D superresolution-
dc.typeArticle-
dc.identifier.hkuros344748-
dc.identifier.eissn1522-2594-
dc.identifier.issnl0740-3194-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats