Deep learning-based reconstruction on intensity-inhomogeneous diffusion magnetic resonance imaging
Zaimin Zhu
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
Contribution: Conceptualization (equal), Data curation (equal), Methodology (equal), Project administration (equal), Software (equal), Validation (equal), Visualization (equal), Writing - original draft (equal), Writing - review & editing (equal)
Search for more papers by this authorHe Wang
Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai, China
Contribution: Project administration (equal), Writing - review & editing (supporting)
Search for more papers by this authorYong Liu
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
Queen Mary School Hainan, Beijing University of Posts and Telecommunications, Lingshui, Hainan, China
Contribution: Project administration (equal), Supervision (equal)
Search for more papers by this authorCorresponding Author
Fangrong Zong
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
Queen Mary School Hainan, Beijing University of Posts and Telecommunications, Lingshui, Hainan, China
Correspondence
Fangrong Zong.
Email: [email protected]
Contribution: Conceptualization (equal), Data curation (equal), Funding acquisition (equal), Project administration (equal), Supervision (equal), Writing - review & editing (equal)
Search for more papers by this authorZaimin Zhu
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
Contribution: Conceptualization (equal), Data curation (equal), Methodology (equal), Project administration (equal), Software (equal), Validation (equal), Visualization (equal), Writing - original draft (equal), Writing - review & editing (equal)
Search for more papers by this authorHe Wang
Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai, China
Contribution: Project administration (equal), Writing - review & editing (supporting)
Search for more papers by this authorYong Liu
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
Queen Mary School Hainan, Beijing University of Posts and Telecommunications, Lingshui, Hainan, China
Contribution: Project administration (equal), Supervision (equal)
Search for more papers by this authorCorresponding Author
Fangrong Zong
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
Queen Mary School Hainan, Beijing University of Posts and Telecommunications, Lingshui, Hainan, China
Correspondence
Fangrong Zong.
Email: [email protected]
Contribution: Conceptualization (equal), Data curation (equal), Funding acquisition (equal), Project administration (equal), Supervision (equal), Writing - review & editing (equal)
Search for more papers by this authorAbstract
Background
Ultra high field diffusion magnetic resonance imaging (dMRI) provides diffusion-weighted (DW) images with a high signal-to-noise ratio, but increases inhomogeneity, which affects the accuracy of dMRI metric reconstruction. Current methods for correcting inhomogeneity rarely consider the accuracy of the reconstructed dMRI metrics. Deep learning models for reconstructing metrics from dMRI signals typically assume that DW images have a homogeneous intensity. To address these challenges, we propose a deep learning model capable of directly reconstructing high-accuracy dMRI metric maps from inhomogeneous DW images.
Methods
An attention-based q-space inhomogeneity-resistant reconstruction network (qIRR-Net) is proposed for the voxel-wise reconstruction of diffusion tensor imaging and diffusion kurtosis imaging metrics. A training procedure based on data augmentation and consistency loss is introduced to ensure that the reconstruction results of qIRR-Net are not affected by signal inhomogeneity. The 3T and 7T dMRI data from the Human Connectome Project are used for model training, testing, and evaluation.
Results
On the 3T dMRI data with simulated inhomogeneity, qIRR-Net improves the peak signal-to-noise ratio by 5.39 and the structural similarity index measure by 0.18 compared with weighted linear least-squares fitting. On the 7T dMRI data, the metric maps reconstructed by qIRR-Net not only exhibit clearer tissue structures but also demonstrate greater stability compared with the weighted linear least-squares results.
Conclusions
The proposed qIRR-Net enables the accurate reconstruction of dMRI metrics from inhomogeneous DW images. This approach could potentially be expanded to obtain multiple artifact-free metric maps from ultrahigh field dMRI for neuroscience research and neurology applications.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Open Research
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available in WU-Minn HCP Data at https://db.humanconnectome.org. The code and the trained model are available at https://github.com/AI4DMR/qIRR-Net.
REFERENCES
- 1Baliyan V, Das CJ, Sharma R, Gupta AK. Diffusion weighted imaging: technique and applications. World J Radiol. 2016; 8(9): 785–798. https://doi.org/10.4329/wjr.v8.i9.785
- 2Mori S, Barker PB. Diffusion magnetic resonance imaging: its principle and applications. Anat Rec. 1999; 257(3): 102–109. https://doi.org/10.1002/(sici)1097-0185(19990615)257:3<102:aid-ar7>3.0.co;2-6
10.1002/(SICI)1097-0185(19990615)257:3<102::AID-AR7>3.0.CO;2-6 CAS PubMed Web of Science® Google Scholar
- 3Song X, Zhou F, Frangi AF, Cao J, Xiao X, Lei Y, et al. Multicenter and multichannel pooling GCN for early AD diagnosis based on dual-modality fused brain network. IEEE Trans Med Imag. 2022; 42(2): 354–367. https://doi.org/10.1109/TMI.2022.3187141
- 4Song X, Li J, Qian X. Diagnosis of glioblastoma multiforme progression via interpretable structure-constrained graph neural networks. IEEE Trans Med Imag. 2023; 42(2): 380–390. https://doi.org/10.1109/TMI.2022.3202037
- 5Chenevert TL, Stegman LD, Taylor JM, Robertson PL, Greenberg HS, Rehemtulla A, et al. Diffusion magnetic resonance imaging: an early surrogate marker of therapeutic efficacy in brain tumors. J Natl Cancer Inst. 2000; 92(24): 2029–2036. https://doi.org/10.1093/jnci/92.24.2029
- 6Basser PJ, Pajevic S, Pierpaoli C, Duda J, Aldroubi A. In vivo fiber tractography using DT-MRI data. Magn Reson Med. 2000; 44(4): 625–632. https://doi.org/10.1002/1522-2594(200010)44:4<625::AID-MRM17>3.0.CO;2-O
- 7Albers GW. Diffusion-weighted MRI for evaluation of acute stroke. Neurology. 1998; 51(Suppl 3): S47–S49. https://doi.org/10.1212/wnl.51.3_suppl_3.s47
- 8Basser PJ, Mattiello J, LeBihan D. MR diffusion tensor spectroscopy and imaging. Biophys J. 1994; 66(1): 259–267. https://doi.org/10.1016/s0006-3495(94)80775-1
- 9Jensen JH, Helpern JA, Ramani A, Lu H, Kaczynski K. Diffusional kurtosis imaging: the quantification of non-Gaussian water diffusion by means of magnetic resonance imaging. Magn Reson Med. 2005; 53(6): 1432–1440. https://doi.org/10.1002/mrm.20508
- 10Uğurbil K. Imaging at ultrahigh magnetic fields: history, challenges, and solutions. Neuroimage. 2018; 168: 7–32. https://doi.org/10.1016/j.neuroimage.2017.07.007
- 11Gallichan D. Diffusion MRI of the human brain at ultra-high field (UHF):a review. Neuroimage. 2018; 168: 172–180. https://doi.org/10.1016/j.neuroimage.2017.04.037
- 12Feinberg DA, Beckett AJS, Vu AT, Stockmann J, Huber L, Ma S, et al. Next-generation MRI scanner designed for ultra-high-resolution human brain imaging at 7 Tesla. Nat Methods. 2023; 20(12): 2048–2057. https://doi.org/10.1038/s41592-023-02068-7
- 13Truong TK, Clymer BD, Chakeres DW, Schmalbrock P. Three-dimensional numerical simulations of susceptibility-induced magnetic field inhomogeneities in the human head. Magn Reson Imaging. 2002; 20(10): 759–770. https://doi.org/10.1016/S0730-725X(02)00601-X
- 14Ibrahim TS, Abduljalil AM, Baertlein BA, Lee R, Robitaill PM. Analysis of B1 field profiles and SAR values for multi-strut transverse electromagnetic RF coils in high field MRI applications. Phys Med Biol. 2001; 46(10): 2545–2555. https://doi.org/10.1088/0031-9155/46/10/303
- 15Sled JG, Zijdenbos AP, Evans AC. A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans Med Imag. 1998; 17(1): 87–97. https://doi.org/10.1109/42.668698
- 16Tustison NJ, Avants BB, Cook PA, Zheng Y, Egan A, Yushkevich PA, et al. N4ITK: improved N3 bias correction. IEEE Trans Med Imag. 2010; 29(6): 1310–1320. https://doi.org/10.1109/TMI.2010.2046908
- 17Mikheev A, Bokacheva L, Yang H, Sobel J, Fernandez-Granda C, Chandarana H, et al. Retrospective MRI non-uniformity correction: quantitative assessment of two methods. In: Poster presented at: ISMRM and SMRT virtual conference and exhibition; 2020.
- 18Simkó A, Löfstedt T, Garpebring A, Nyholm T, Jonsson J. MRI bias field correction with an implicitly trained CNN. In: International conference on medical imaging with deep learning. PMLR; 2022. https://doi.org/10.5281/zenodo.3749526
- 19Chuang KH, Wu PH, Li Z, Fan KH, Weng JC. Deep learning network for integrated coil inhomogeneity correction and brain extraction of mixed MRI data. Sci Rep. 2022; 12(1):8578. https://doi.org/10.1038/s41598-022-12587-6
- 20Goldfryd T, Gordon S, Raviv TR. Deep semi-supervised bias field correction of mr images. In: 2021 IEEE 18th international symposium on biomedical imaging (ISBI). Nice; 2021. p. 1836–1840. https://doi.org/10.1109/ISBI48211.2021.9433889
10.1109/ISBI48211.2021.9433889 Google Scholar
- 21Venkatesh V, Sharma N, Singh M. Intensity inhomogeneity correction of MRI images using InhomoNet. Comput Med Imag Graph. 2020; 84:101748. https://doi.org/10.1016/j.compmedimag.2020.101748
- 22Gaillochet M, Tezcan KC, Konukoglu E. Joint reconstruction and bias field correction for undersampled MR imaging. In: International conference on medical image computing and computer-assisted intervention. Springer; 2020. https://doi.org/10.1007/978-3-030-59713-9_5
10.1007/978-3-030-59713-9_5 Google Scholar
- 23Dai X, Lei Y, Liu Y, Wang T, Ren L, Curran WJ, et al. Intensity non-uniformity correction in MR imaging using residual cycle generative adversarial network. Phys Med Biol. 2020; 65(21):215025. https://doi.org/10.1088/1361-6560/abb31f
- 24Wan F, Smedby Ö, Wang C. Simultaneous MR knee image segmentation and bias field correction using deep learning and partial convolution. In: Medical imaging 2019: image processing. San Diego; 2019. https://doi.org/10.1117/12.2512950
10.1117/12.2512950 Google Scholar
- 25Harrevelt SD, Meliado EFM, van Lier ALHMW, Reesink D, Meijer RP, Pluim JPW, et al. Deep learning based correction of RF field induced inhomogeneities for T2w prostate imaging at 7 T. NMR Biomed. 2023; 36(12):e5019. https://doi.org/10.1002/nbm.5019
- 26Zheng T, Yan G, Li H, Zheng W, Shi W, Zhang Y, et al. A microstructure estimation Transformer inspired by sparse representation for diffusion MRI. Med Image Anal. 2023; 86:102788. https://doi.org/10.1016/j.media.2023.102788
- 27Ye C, Li Y, Zeng X. An improved deep network for tissue microstructure estimation with uncertainty quantification. Med Image Anal. 2020; 61:101650. https://doi.org/10.1016/j.media.2020.101650
- 28Chen G, Hong Y, Zhang Y, Kim J, Huynh KM, Ma J, et al. Estimating tissue microstructure with undersampled diffusion data via graph convolutional neural networks. Med Image Comput Comput Assist Interv. 2020; 12267: 280–290. https://doi.org/10.1007/978-3-030-59728-3_28
- 29Golkov V, Dosovitskiy A, Sperl JI, Menzel MI, Czisch M, Samann P, et al. Q-space deep learning: twelve-fold shorter and model-free diffusion MRI scans. IEEE Trans Med Imag. 2016; 35(5): 1344–1351. https://doi.org/10.1109/tmi.2016.2551324
- 30Ewert C, Kügler D, Stirnberg R, Koch A, Yendiki A, Reuter M. Geometric deep learning for diffusion MRI signal reconstruction with continuous samplings (DISCUS). Imag Neurosci. 2024; 2: 1–18. https://doi.org/10.1162/imag_a_00121
10.1162/imag_a_00121 Google Scholar
- 31Park J, Jung W, Choi EJ, Oh SH, Jang J, Shin D, et al. DIFFnet: diffusion parameter mapping network generalized for input diffusion gradient schemes and b-value. IEEE Trans Med Imag. 2022; 41(2): 491–499. https://doi.org/10.1109/tmi.2021.3116298
- 32Li Z, Gong T, Lin Z, He H, Tong Q, Li C, et al. Fast and robust diffusion kurtosis parametric mapping using a three-dimensional convolutional neural network. IEEE Access. 2019; 7: 71398–71411. https://doi.org/10.1109/ACCESS.2019.2919241
- 33Gibbons EK, Hodgson KK, Chaudhari AS, Richards LG, Majersik JJ, Adluru G, et al. Simultaneous NODDI and GFA parameter map generation from subsampled q-space imaging using deep learning. Magn Reson Med. 2019; 81(4): 2399–2411. https://doi.org/10.1002/mrm.27568
- 34Ye C. Tissue microstructure estimation using a deep network inspired by a dictionary-based framework. Med Image Anal. 2017; 42: 288–299. https://doi.org/10.1016/j.media.2017.09.001
- 35Van Essen DC, Smith SM, Barch DM, Behrens TEJ, Yacoub E, Ugurbil K, et al. The Wu-minn human connectome project: an overview. Neuroimage. 2013; 80: 62–79. https://doi.org/10.1016/j.neuroimage.2013.05.041
- 36Veraart J, Sijbers J, Sunaert S, Leemans A, Jeurissen B. Weighted linear least squares estimation of diffusion MRI parameters: strengths, limitations, and pitfalls. Neuroimage. 2013; 81: 335–346. https://doi.org/10.1016/j.neuroimage.2013.05.028
- 37Chung S, Lu Y, Henry RG. Comparison of bootstrap approaches for estimation of uncertainties of DTI parameters. Neuroimage. 2006; 33(2): 531–541. https://doi.org/10.1016/j.neuroimage.2006.07.001
- 38Garyfallidis E, Brett M, Amirbekian B, Rokem A, van der Walt S, Descoteaux M, et al. Dipy, a library for the analysis of diffusion MRI data. Front Neuroinf. 2014; 8:8. https://doi.org/10.3389/fninf.2014.00008
- 39Glasser MF, Sotiropoulos SN, Wilson JA, Coalson TS, Fischl B, Andersson JL, et al. The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage. 2013; 80: 105–124. https://doi.org/10.1016/j.neuroimage.2013.04.127
- 40Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez A N, et al. Attention is all you need. Adv Neural Inf Process Syst. 2017; 30: 1–15. https://doi.org/10.48550/arXiv.1706.03762
10.48550/arXiv.1706.03762 Google Scholar
- 41Borkowski K, Krzyżak AT. Assessment of the systematic errors caused by diffusion gradient inhomogeneity in DTI-computer simulations. NMR Biomed. 2019; 32(11):e4130. https://doi.org/10.1002/nbm.4130
- 42Yeh FC, Verstynen TD, Wang Y, Fernández-Miranda JC, Tseng WYI. Deterministic diffusion fiber tracking improved by quantitative anisotropy. PLoS One. 2013; 8(11):e80713. https://doi.org/10.1371/journal.pone.0080713
- 43Deissler RJ, Wu Y, Martens MA. Dependence of Brownian and Néel relaxation times on magnetic field strength. Med Phys. 2014; 41(1):012301. https://doi.org/10.1118/1.4837216
- 44Gomori JM, Grossman RI, Yu-Ip C, Asakura T. NMR relaxation times of blood: dependence on field strength, oxidation state, and cell integrity. J Comput Assist Tomogr. 1987; 11(4): 684–690. https://doi.org/10.1097/00004728-198707000-00025
- 45Zhang Y, Dai Y, Zeng M. Higher field reduced-FOV diffusion-weighted-imaging for abdominal imaging at 5.0 tesla: image quality evaluation compared with 3.0 tesla. In: ISMRM annual meeting. Toronto; 2023. https://doi.org/10.58530/2023/3808
- 46Shi Z, Zhao X, Zhu S, Miao X, Zhang Y, Han S, et al. Time-of-flight intracranial MRA at 3 T versus 5 T versus 7 T: visualization of distal small cerebral arteries. Radiology. 2023; 306(1): 207–217. https://doi.org/10.1148/radiol.220114
- 47Jiang Z, Sun W, Xu D, Yu H, Mei H, Song X, et al. Stability and repeatability of diffusion-weighted imaging (DWI) of normal pancreas on 5.0 Tesla magnetic resonance imaging (MRI). Sci Rep. 2023; 13(1):11954. https://doi.org/10.1038/s41598-023-38360-x
- 48Zhang Y, Yang C, Liang L, Shi Z, Zhu S, Chen C, et al. Preliminary experience of 5.0 T higher field abdominal diffusion-weighted MRI: agreement of apparent diffusion coefficient with 3.0 T imaging. J Magn Reson Imag. 2022; 56(4): 1009–1017. https://doi.org/10.1002/jmri.28097
- 49Wei Z, Chen Q, Han S, Zhang S, Zhang N, Zhang L, et al. 5T magnetic resonance imaging: radio frequency hardware and initial brain imaging. Quant Imag Med Surg. 2023; 13(5): 3222–3240. https://doi.org/10.21037/qims-22-945
