Phase-constrained reconstruction method with compressed sensing for multi-parametric quantitative magnetic resonance imaging. (February 2023)

• Record Type: Journal Article
• Authors: Guo, Hongyu; Zhou, Jianing; Liu, Peng; Chen, Hong
• Is Part Of: Biomedical signal processing and control. Volume 80:Part 2(2023)
• Publication Date: 2023-02
• Subjects: Multi-contrast -- Phase-constrained reconstruction -- Compressed sensing -- Quantitative magnetic resonance imaging; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104383
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Phase-constrained reconstruction method combined with compressed sensing and parallel imaging. Phase consistency between data from different flip angles in STAGE. Primal–dual algorithm to solve the proposed model with multiple regularization terms. Accelerate STAGE. Abstract: Purpose: In magnetic resonance imaging (MRI), STAGE method is a novel quantitative imaging method to generate multi-parametric quantitative images during single one scan. Due to the use of multiple echo sequences with double flip angles, current acquisition methods consume a lot of time, which limits the clinical application of these methods. To shorten the acquisition time of the STAGE method, a phase-constrained CS reconstruction method combined with parallel imaging (PI) is proposed in this paper. Methods: The phase consistency between the acquisition data from different flip angles (FA) is exploited to reconstruct STAGE sequences. The phase from the first flip acquisition is used as the constraint term in the CS model to reconstruct the acquisition data from the second flip angle. To solve the optimization problem with multiple regularization in the CS model, the primal–dual algorithm is applied to the optimization problem. In vivo data from 10 volunteers are reconstructed, and the reconstruction results are compared with those of the conventional l 1 -iterative self-consistent parallel imaging reconstruction from arbitrary k-space ( l 1 -SPIRiT) method. Also, quantitative comparisons are performed with multiple quantitative matrics including structural similarity index measure (SSIM), peak signal-to-noise ratio (pSNR) for the contrast image, normalized root-mean-square error (NRMS), and linear relationship value. Results: The experimental results demonstrate that the proposed method outperforms the conventional l 1 -SPIRiT method under different acceleration factors. Meanwhile, the proposed method achieves better SSIM and higher pSNR than the conventional l 1 -SPIRiT for T1 weighted (T1W) images. Besides, the quantitative mapping obtained by the proposed method under different acceleration factors have better consistency with the reference one. Moreover, the proposed method enables further acceleration of the image acquisition of STAGE sequences. Conclusion: The proposed phase-constrained reconstruction method with compressed sensing accelerates STAGE without compromising the image quality, which enables higher acceleration than the conventional method.