B 1 mapping using pre‐learned subspaces for quantitative brain imaging
Magnetic Resonance in Medicine · 2023
DOI: 10.1002/mrm.29764 ↗Kurz gefasst
Fachartikel in der Zeitschrift Magnetic Resonance in Medicine, 2023. Verfasst von Tianxiao Zhang und 8 weiteren Autor:innen. Beteiligte Einrichtung: School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China. Hauptthemen laut MEDLINE: Magnetic Resonance Imaging, Brain Neoplasms. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 60 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning.
Abstract (Originalsprache)
Purpose
To develop a machine learning‐based method for estimation of both transmitter and receiver B 1 fields desired for correction of the B 1 inhomogeneity effects in quantitative brain imaging. Theory and Methods A subspace model‐based machine learning method was proposed for estimation of B 1t and B 1r fields. Probabilistic subspace models were used to capture scan‐dependent variations in the B 1 fields; the subspace basis and coefficient distributions were learned from pre‐scanned training data. Estimation of the B 1 fields for new experimental data was achieved by solving a linear optimization problem with prior distribution constraints. We evaluated the performance of the proposed method for B 1 inhomogeneity correction in quantitative brain imaging scenarios, including T 1 and proton density (PD) mapping from variable‐flip‐angle spoiled gradient‐echo (SPGR) data as well as neurometabolic mapping from MRSI data, using phantom, healthy subject and brain tumor patient data.
Results
In both phantom and healthy subject data, the proposed method produced high‐quality B 1 maps. B 1 correction on SPGR data using the estimated B 1 maps produced significantly improved T 1 and PD maps. In brain tumor patients, the proposed method produced more accurate and robust B 1 estimation and correction results than conventional methods. The B 1 maps were also applied to MRSI data from tumor patients and produced improved neurometabolite maps, with better separation between pathological and normal tissues.
Conclusion
This work presents a novel method to estimate B 1 variations using probabilistic subspace models and machine learning. The proposed method may make correction of B 1 inhomogeneity effects more robust in practical applications.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China.
- Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, Illinois, USA.
- Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, USA.
- Siemens Medical Solutions USA, Inc., Urbana, Illinois, USA.
ORCID-Profile: Tianxiao Zhang, Yibo Zhao, Yudu Li, Rong Guo, Jie Luo, Yao Li
Förderung
- National Natural Science Foundation of China
- Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning
Zitieren
APA
Zhang, T., Zhao, Y., Jin, W., Li, Y., Guo, R., Ke, Z., Luo, J., Li, Y., & Liang, Z. P. (2023). B 1 mapping using pre‐learned subspaces for quantitative brain imaging. Magnetic Resonance in Medicine. https://doi.org/10.1002/mrm.29764
Vancouver
Zhang T, Zhao Y, Jin W, Li Y, Guo R, Ke Z, et al. B 1 mapping using pre‐learned subspaces for quantitative brain imaging. Magnetic Resonance in Medicine. 2023. doi:10.1002/mrm.29764
BibTeX
@article{Zhang2023,
author = {Zhang, Tianxiao and Zhao, Yibo and Jin, Wen and Li, Yudu and Guo, Rong and Ke, Ziwen and Luo, Jie and Li, Yao and Liang, Zhi-Pei},
title = {B 1 mapping using pre‐learned subspaces for quantitative brain imaging},
journal = {Magnetic Resonance in Medicine},
year = {2023},
doi = {10.1002/mrm.29764}
}
EUVIMED