Harmonizing three-dimensional MRI using pseudo-warping field guided GAN

Jiaying Lin; Zhuoshuo Li; Youbing Zeng; Xiaobo Liu; Liang Li; Neda Jahanshad; Xinting Ge; Dan Zhang; Minhua Lu; Mengting Liu; Alzheimer’s Disease Neuroimaging Initiative

doi:10.1016/j.neuroimage.2024.120635

Harmonizing three-dimensional MRI using pseudo-warping field guided GAN

Neuroimage. 2024 May 8:295:120635. doi: 10.1016/j.neuroimage.2024.120635. Online ahead of print.

Authors

Affiliations

¹ Department of Biomedical Engineering, Sun Yat-sen University, Shenzhen 518107, China. Electronic address: linjy277@mail2.sysu.edu.cn.
² Department of Biomedical Engineering, Sun Yat-sen University, Shenzhen 518107, China. Electronic address: lizhsh23@mail2.sysu.edu.cn.
³ Department of Biomedical Engineering, Sun Yat-sen University, Shenzhen 518107, China. Electronic address: zengyb6@mail2.sysu.edu.cn.
⁴ Department of Biomedical Engineering, Sun Yat-sen University, Shenzhen 518107, China. Electronic address: liuxb26@mail2.sysu.edu.cn.
⁵ Genuine Digital Technology Co., Ltd., Xi'an, China. Electronic address: liliang@genuai.com.
⁶ Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA. Electronic address: njahansh@usc.edu.
⁷ School of Information Science and Engineering, Shandong Normal University, Shandong 250358, China. Electronic address: xintingge@163.com.
⁸ School of Cyber Science and Engineering, Ningbo University of Technology, Zhejiang 315211, China. Electronic address: danzhang@nbut.edu.cn.
⁹ Department of Biomedical Engineering, Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, Shenzhen University, Shenzhen 518060, China. Electronic address: luminhua@szu.edu.cn.
¹⁰ Department of Biomedical Engineering, Sun Yat-sen University, Shenzhen 518107, China. Electronic address: liumt55@mail.sysu.edu.cn.

PMID: 38729542
DOI: 10.1016/j.neuroimage.2024.120635

Abstract

In pursuit of cultivating automated models for magnetic resonance imaging (MRI) to aid in diagnostics, an escalating demand for extensive, multisite, and heterogeneous brain imaging datasets has emerged. This potentially introduces biased outcomes when directly applied for subsequent analysis. Researchers have endeavored to address this issue by pursuing the harmonization of MRIs. However, most existing image-based harmonization methods for MRI are tailored for 2D slices, which may introduce inter-slice variations when they are combined into a 3D volume. In this study, we aim to resolve inconsistencies between slices by introducing a pseudo-warping field. This field is created randomly and utilized to transform a slice into an artificially warped subsequent slice. The objective of this pseudo-warping field is to ensure that generators can consistently harmonize adjacent slices to another domain, without being affected by the varying content present in different slices. Furthermore, we construct unsupervised spatial and recycle loss to enhance the spatial accuracy and slice-wise consistency across the 3D images. The results demonstrate that our model effectively mitigates inter-slice variations and successfully preserves the anatomical details of the images during the harmonization process. Compared to generative harmonization models that employ 3D operators, our model exhibits greater computational efficiency and flexibility.

Keywords: Brain MRI; GAN; Pseudo-warping field; Slice-wise 3D MRI harmonization.