View in EDS HTML Full Text PDF Full Text

Generalizable, sequence‐invariant deep learning image reconstruction for subspace‐constrained quantitative MRI.

Saved in:
Bibliographic Details
Title: Generalizable, sequence‐invariant deep learning image reconstruction for subspace‐constrained quantitative MRI.
Authors: Hu, Zheyuan1,2,3 (AUTHOR), Chen, Zihao1,2,3 (AUTHOR), Cao, Tianle1,2,3 (AUTHOR), Lee, Hsu‐Lei3 (AUTHOR), Xie, Yibin3 (AUTHOR), Li, Debiao2,3 (AUTHOR), Christodoulou, Anthony G.1,2,3 (AUTHOR) achristodoulou@mednet.ucla.edu
Source: Magnetic Resonance in Medicine. Jul2025, Vol. 94 Issue 1, p89-104. 16p.
Subjects: Standard deviations, Deep learning, Cardiac magnetic resonance imaging, Image reconstruction, Magnetic resonance imaging
Abstract: Purpose: To develop a deep subspace learning network that can function across different pulse sequences. Methods: A contrast‐invariant component‐by‐component (CBC) network structure was developed and compared against previously reported spatiotemporal multicomponent (MC) structure for reconstructing MR Multitasking images. A total of 130, 167, and 16 subjects were imaged using T1, T1‐T2, and T1‐T2‐ T2*$$ {\mathrm{T}}_2^{\ast } $$‐fat fraction (FF) mapping sequences, respectively. We compared CBC and MC networks in matched‐sequence experiments (same sequence for training and testing), then examined their cross‐sequence performance and generalizability by unmatched‐sequence experiments (different sequences for training and testing). A "universal" CBC network was also evaluated using mixed‐sequence training (combining data from all three sequences). Evaluation metrics included image normalized root mean squared error and Bland–Altman analyses of end‐diastolic maps, both versus iteratively reconstructed references. Results: The proposed CBC showed significantly better normalized root mean squared error than MC in both matched‐sequence and unmatched‐sequence experiments (p < 0.001), fewer structural details in quantitative error maps, and tighter limits of agreement. CBC was more generalizable than MC (smaller performance loss; p = 0.006 in T1 and p < 0.001 in T1‐T2 from matched‐sequence testing to unmatched‐sequence testing) and additionally allowed training of a single universal network to reconstruct images from any of the three pulse sequences. The mixed‐sequence CBC network performed similarly to matched‐sequence CBC in T1 (p = 0.178) and T1‐T2 (p = 0121), where training data were plentiful, and performed better in T1‐T2‐T2*$$ {\mathrm{T}}_2^{\ast } $$‐FF (p < 0.001) where training data were scarce. Conclusion: Contrast‐invariant learning of spatial features rather than spatiotemporal features improves performance and generalizability, addresses data scarcity, and offers a pathway to universal supervised deep subspace learning. [ABSTRACT FROM AUTHOR]
Copyright of Magnetic Resonance in Medicine is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
Text:
  Availability: 1
Header DbId: egs
DbLabel: Engineering Source
An: 184713961
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Generalizable, sequence‐invariant deep learning image reconstruction for subspace‐constrained quantitative MRI.
– Name: Author
  Label: Authors
  Group: Au
  Data: &lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Hu%2C+Zheyuan%22&quot;&gt;Hu, Zheyuan&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Chen%2C+Zihao%22&quot;&gt;Chen, Zihao&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Cao%2C+Tianle%22&quot;&gt;Cao, Tianle&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Lee%2C+Hsu‐Lei%22&quot;&gt;Lee, Hsu‐Lei&lt;/searchLink&gt;&lt;relatesTo&gt;3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Xie%2C+Yibin%22&quot;&gt;Xie, Yibin&lt;/searchLink&gt;&lt;relatesTo&gt;3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Li%2C+Debiao%22&quot;&gt;Li, Debiao&lt;/searchLink&gt;&lt;relatesTo&gt;2,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Christodoulou%2C+Anthony+G%2E%22&quot;&gt;Christodoulou, Anthony G.&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3&lt;/relatesTo&gt; (AUTHOR)&lt;i&gt; achristodoulou@mednet.ucla.edu&lt;/i&gt;
– Name: TitleSource
  Label: Source
  Group: Src
  Data: &lt;searchLink fieldCode=&quot;JN&quot; term=&quot;%22Magnetic+Resonance+in+Medicine%22&quot;&gt;Magnetic Resonance in Medicine&lt;/searchLink&gt;. Jul2025, Vol. 94 Issue 1, p89-104. 16p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: &lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Standard+deviations%22&quot;&gt;Standard deviations&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Deep+learning%22&quot;&gt;Deep learning&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Cardiac+magnetic+resonance+imaging%22&quot;&gt;Cardiac magnetic resonance imaging&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Image+reconstruction%22&quot;&gt;Image reconstruction&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Magnetic+resonance+imaging%22&quot;&gt;Magnetic resonance imaging&lt;/searchLink&gt;
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Purpose: To develop a deep subspace learning network that can function across different pulse sequences. Methods: A contrast‐invariant component‐by‐component (CBC) network structure was developed and compared against previously reported spatiotemporal multicomponent (MC) structure for reconstructing MR Multitasking images. A total of 130, 167, and 16 subjects were imaged using T1, T1‐T2, and T1‐T2‐ T2*$$ {\mathrm{T}}_2^{\ast } $$‐fat fraction (FF) mapping sequences, respectively. We compared CBC and MC networks in matched‐sequence experiments (same sequence for training and testing), then examined their cross‐sequence performance and generalizability by unmatched‐sequence experiments (different sequences for training and testing). A &quot;universal&quot; CBC network was also evaluated using mixed‐sequence training (combining data from all three sequences). Evaluation metrics included image normalized root mean squared error and Bland–Altman analyses of end‐diastolic maps, both versus iteratively reconstructed references. Results: The proposed CBC showed significantly better normalized root mean squared error than MC in both matched‐sequence and unmatched‐sequence experiments (p &lt; 0.001), fewer structural details in quantitative error maps, and tighter limits of agreement. CBC was more generalizable than MC (smaller performance loss; p = 0.006 in T1 and p &lt; 0.001 in T1‐T2 from matched‐sequence testing to unmatched‐sequence testing) and additionally allowed training of a single universal network to reconstruct images from any of the three pulse sequences. The mixed‐sequence CBC network performed similarly to matched‐sequence CBC in T1 (p = 0.178) and T1‐T2 (p = 0121), where training data were plentiful, and performed better in T1‐T2‐T2*$$ {\mathrm{T}}_2^{\ast } $$‐FF (p &lt; 0.001) where training data were scarce. Conclusion: Contrast‐invariant learning of spatial features rather than spatiotemporal features improves performance and generalizability, addresses data scarcity, and offers a pathway to universal supervised deep subspace learning. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: &lt;i&gt;Copyright of Magnetic Resonance in Medicine is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder&#39;s express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.&lt;/i&gt; (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=184713961
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1002/mrm.30433
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 16
        StartPage: 89
    Subjects:
      – SubjectFull: Standard deviations
        Type: general
      – SubjectFull: Deep learning
        Type: general
      – SubjectFull: Cardiac magnetic resonance imaging
        Type: general
      – SubjectFull: Image reconstruction
        Type: general
      – SubjectFull: Magnetic resonance imaging
        Type: general
    Titles:
      – TitleFull: Generalizable, sequence‐invariant deep learning image reconstruction for subspace‐constrained quantitative MRI.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Hu, Zheyuan
      – PersonEntity:
          Name:
            NameFull: Chen, Zihao
      – PersonEntity:
          Name:
            NameFull: Cao, Tianle
      – PersonEntity:
          Name:
            NameFull: Lee, Hsu‐Lei
      – PersonEntity:
          Name:
            NameFull: Xie, Yibin
      – PersonEntity:
          Name:
            NameFull: Li, Debiao
      – PersonEntity:
          Name:
            NameFull: Christodoulou, Anthony G.
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 07
              Text: Jul2025
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 07403194
          Numbering:
            – Type: volume
              Value: 94
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Magnetic Resonance in Medicine
              Type: main
ResultId 1