EUVIMEDEUVIMED entdecken
Open Access

An untrained deep learning method for reconstructing dynamic MR images from accelerated model‐based data

Kalina P Slavkova iD, Julie C DiCarlo iD, Viraj Wadhwa, Sidharth Kumar, Chengyue Wu iD, John Virostko, Thomas E Yankeelov, Jonathan I Tamir iD

Magnetic Resonance in Medicine · 2023-04 · 89 (4) · 1617-1633

DOI: 10.1002/mrm.29547 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Magnetic Resonance in Medicine, online erschienen am 05.12.2022. Verfasst von Kalina P Slavkova und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Physics, The University of Texas, Austin, Texas, USA. Hauptthemen laut MEDLINE: Image Processing, Computer-Assisted, Deep Learning. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 66 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NCI NIH HHS, NIBIB NIH HHS, National Institutes of Health.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

To implement physics‐based regularization as a stopping condition in tuning an untrained deep neural network for reconstructing MR images from accelerated data.

Methods

The ConvDecoder (CD) neural network was trained with a physics‐based regularization term incorporating the spoiled gradient echo equation that describes variable‐flip angle data. Fully‐sampled variable‐flip angle k‐space data were retrospectively accelerated by factors of R = {8, 12, 18, 36} and reconstructed with CD, CD with the proposed regularization (CD + r ), locally low‐rank (LR) reconstruction, and compressed sensing with L1‐wavelet regularization (L1). Final images from CD + r training were evaluated at the “ argmin ” of the regularization loss; whereas the CD, LR, and L1 reconstructions were chosen optimally based on ground truth data. The performance measures used were the normalized RMS error, the concordance correlation coefficient, and the structural similarity index.

Results

The CD + r reconstructions, chosen using the stopping condition, yielded structural similarity indexs that were similar to the CD ( p = 0.47) and LR structural similarity indexs ( p = 0.95) across R and that were significantly higher than the L1 structural similarity indexs ( p = 0.04). The concordance correlation coefficient values for the CD + r T 1 maps across all R and subjects were greater than those corresponding to the L1 ( p = 0.15) and LR ( p = 0.13) T 1 maps, respectively. For R ≥ 12 (≤4.2 min scan time), L1 and LR T 1 maps exhibit a loss of spatially refined details compared to CD + r .

Conclusion

The use of an untrained neural network together with a physics‐based regularization loss shows promise as a measure for determining the optimal stopping point in training without relying on fully‐sampled ground truth data.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Image Processing, Computer-AssistedDeep LearningHumansRetrospective StudiesMagnetic Resonance ImagingNeural Networks, Computer

Stichwörter der Autor:innen

image reconstructioninverse problemphysics-guided deep learningquantitative MRI

Autor:innen und Institutionen

  • Department of Physics, The University of Texas, Austin, Texas, USA.
  • The Oden Institute for Computational Engineering and Sciences, The University of Texas, Austin, Texas, USA.
  • Livestrong Cancer Institutes, Dell Medical School, The University of Texas, Austin, Texas, USA.
  • Chandra Family Department of Electrical and Computer Engineering, The University of Texas, Austin, Texas, USA.
  • Department of Diagnostic Medicine, Dell Medical School, The University of Texas, Austin, Texas, USA.
  • Department of Oncology, Dell Medical School, The University of Texas, Austin, Texas, USA.

ORCID-Profile: Kalina P Slavkova, Julie C DiCarlo, Chengyue Wu, Jonathan I Tamir

Förderung

  • NCI NIH HHS
  • NIBIB NIH HHS
  • National Institutes of Health
  • American Cancer Society
  • Cancer Prevention and Research Institute of Texas

Zitieren

APA

Slavkova, K. P., DiCarlo, J. C., Wadhwa, V., Kumar, S., Wu, C., Virostko, J., Yankeelov, T. E., & Tamir, J. I. (2023). An untrained deep learning method for reconstructing dynamic MR images from accelerated model‐based data. Magnetic Resonance in Medicine, 89(4), 1617-1633. https://doi.org/10.1002/mrm.29547

Vancouver

Slavkova KP, DiCarlo JC, Wadhwa V, Kumar S, Wu C, Virostko J, et al. An untrained deep learning method for reconstructing dynamic MR images from accelerated model‐based data. Magnetic Resonance in Medicine. 2023;89(4):1617-1633. doi:10.1002/mrm.29547

BibTeX
@article{Slavkova2023,
  author = {Slavkova, Kalina P and DiCarlo, Julie C and Wadhwa, Viraj and Kumar, Sidharth and Wu, Chengyue and Virostko, John and Yankeelov, Thomas E and Tamir, Jonathan I},
  title = {An untrained deep learning method for reconstructing dynamic MR images from accelerated model‐based data},
  journal = {Magnetic Resonance in Medicine},
  year = {2023},
  volume = {89},
  number = {4},
  pages = {1617--1633},
  doi = {10.1002/mrm.29547}
}
Kostenlos starten →