EUVIMEDOpdag EUVIMED
Open access

DeepEMC ‐ T 2 mapping: Deep learning–enabled T 2 mapping based on echo modulation curve modeling

Haoyang Pei iD, Timothy M Shepherd, Yao Wang, Fang Liu iD, Daniel K Sodickson, Noam Ben-Eliezer iD, Li Feng iD

Magnetic Resonance in Medicine · 2024

DOI: 10.1002/mrm.30239 ↗

At a glance

Journal article in Magnetic Resonance in Medicine, 2024. Written by Haoyang Pei and 6 co-authors. Institution: Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, New Yor…. Major MEDLINE topics: Deep Learning, Image Processing, Computer-Assisted, Algorithms, Magnetic Resonance Imaging. Cited 3 times so far (Crossref); 39 references. The full text is freely available in PubMed Central. Funded by: NIBIB NIH HHS, NIAMS NIH HHS, National Institutes of Health.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Purpose

Echo modulation curve (EMC) modeling enables accurate quantification of T 2 relaxation times in multi‐echo spin‐echo (MESE) imaging. The standard EMC‐T 2 mapping framework, however, requires sufficient echoes and cumbersome pixel‐wise dictionary‐matching steps. This work proposes a deep learning version of EMC‐T 2 mapping, called DeepEMC‐T 2 mapping , to efficiently estimate accurate T 2 maps from fewer echoes.

Methods

DeepEMC‐T 2 mapping was developed using a modified U‐Net to estimate both T 2 and proton density (PD) maps directly from MESE images. The network implements several new features to improve the accuracy of T 2 /PD estimation. A total of 67 MESE datasets acquired in axial orientation were used for network training and evaluation. An additional 57 datasets acquired in coronal orientation with different scan parameters were used to evaluate the generalizability of the framework. The performance of DeepEMC‐T 2 mapping was evaluated in seven experiments.

Results

Compared to the reference, DeepEMC‐T 2 mapping achieved T 2 estimation errors from 1% to 11% and PD estimation errors from 0.4% to 1.5% with ten/seven/five/three echoes, which are more accurate than standard EMC‐T 2 mapping. By incorporating datasets acquired with different scan parameters and orientations for joint training, DeepEMC‐T 2 exhibits robust generalizability across varying imaging protocols. Increasing the echo spacing and including longer echoes improve the accuracy of parameter estimation. The new features proposed in DeepEMC‐T 2 mapping all enabled more accurate T 2 estimation.

Conclusions

DeepEMC‐T 2 mapping enables simplified, efficient, and accurate T 2 quantification directly from MESE images without dictionary matching. Accurate T 2 estimation from fewer echoes allows for increased volumetric coverage and/or higher slice resolution without prolonging total scan times.

Subjects

MeSH terms (MEDLINE)

Deep LearningImage Processing, Computer-AssistedAlgorithmsMagnetic Resonance ImagingHumansBrain

Author keywords

T2 mappingdeep learningecho modulation curvemulti‐echo spin‐echoquantitative MRI

Authors and institutions

  • Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA.
  • Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA.
  • Department of Electrical and Computer Engineering and Department of Biomedical Engineering, New York University Tandon School of Engineering, New York, New York, USA.
  • Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, Massachusetts, USA.
  • Harvard Medical School, Boston, Massachusetts, USA.
  • Department of Biomedical Engineering, Tel Aviv University, Tel Aviv, Israel.

ORCID profiles: Haoyang Pei, Fang Liu, Noam Ben-Eliezer, Li Feng

Funding

  • NIBIB NIH HHS
  • NIAMS NIH HHS
  • National Institutes of Health

Cite

APA

Pei, H., Shepherd, T. M., Wang, Y., Liu, F., Sodickson, D. K., Ben-Eliezer, N., & Feng, L. (2024). DeepEMC ‐ T 2 mapping: Deep learning–enabled T 2 mapping based on echo modulation curve modeling. Magnetic Resonance in Medicine. https://doi.org/10.1002/mrm.30239

Vancouver

Pei H, Shepherd TM, Wang Y, Liu F, Sodickson DK, Ben-Eliezer N, et al. DeepEMC ‐ T 2 mapping: Deep learning–enabled T 2 mapping based on echo modulation curve modeling. Magnetic Resonance in Medicine. 2024. doi:10.1002/mrm.30239

BibTeX
@article{Pei2024,
  author = {Pei, Haoyang and Shepherd, Timothy M and Wang, Yao and Liu, Fang and Sodickson, Daniel K and Ben-Eliezer, Noam and Feng, Li},
  title = {DeepEMC ‐ T 2 mapping: Deep learning–enabled T 2 mapping based on echo modulation curve modeling},
  journal = {Magnetic Resonance in Medicine},
  year = {2024},
  doi = {10.1002/mrm.30239}
}
Start for free →