Share:
Share this content in WeChat
X
Special Focus
Application of deep learning in intravoxel incoherent motion brain magnetic resonance imaging quality
LI Qiongge  Yayan YIN  ZHAO Cheng  QI Zhigang  LU Jie 

DOI:10.12015/issn.1674-8034.2023.05.004.


[Abstract] Objective To explore the clinical implication of deep learning (DL) reconstruction in accelerate intravoxel incoherent motion (IVIM) magnetic resonance imaging (MRI).Materials and Methods Based on 3.0 T MRI system, a total of 40 healthy subjects were collected for accelerated IVIM data acquisition (DL_IVIM) based on DL reconstruction with excitation times (i.e. signal acquisition times) of 1 and conventional image acquisition (ORIG_IVIM) with excitation times of 2. Wilcoxon rank-sum test was used to compare scan time, signal of noise ratio (SNR), non-uniformity index (NUI), quantitative parameters [diffusion coefficient (D value), pseudo-diffusion coefficient (D* value) and perfusion fraction (F value)]. The χ2 test was used to compare the subjective ratings of DL_IVIM and ORGI_IVIM.Results Compared with ORIG_IVIM, the scan time of DL_IVIM was reduced by 23.4%. In addition, SNR (except b=0 s/mm2) and subjective scores of DL_IVIM were significantly higher than those of ORIG_IVIM (P<0.05), NUI (except b=0 s/mm2) were significantly lower than those of ORIG_IVIM (P<0.05), there was no significant difference in D, D* and F values (P>0.05). DL_IVIM score was significantly higher than ORIG_IVIM (χ2=32.81, P<0.001).Conclusions DL reconstruction can significantly improve the image quality of IVIM, ensure the accuracy of quantitative parameters, and shorten the scanning time, which provides valuable reference information for the application of DL reconstruction in clinical IVIM imaging.
[Keywords] intravoxel incoherent motion;deep learning;brain;diffusion weighted imaging;magnetic resonance imaging

LI Qiongge1, 2   Yayan YIN1, 2   ZHAO Cheng1, 2   QI Zhigang1, 2   LU Jie1, 2*  

1 Department of Radiology and Nuclear Medicine, Xuanwu Hospital Capital Medical University, Beijing 100053, China

2 Beijing Key Laboratory of Magnetic Resonance Imaging and Brain Informatics, Beijing 100053, China

Corresponding author: Lu J, E-mail: imaginglu@hotmail.com

Conflicts of interest   None.

Received  2023-01-04
Accepted  2023-04-11
DOI: 10.12015/issn.1674-8034.2023.05.004

[1]
LE BIHAN D. What can we see with IVIM MRI?[J]. Neuroimage, 2019, 187: 56-67. DOI: 10.1016/j.neuroimage.2017.12.062">10.1016/j.neuroimage.2017.12.062">10.1016/j.neuroimage.2017.12.062.
[2]
HU W J, CHEN L H, LIN L J, et al. Three-dimensional amide proton transfer-weighted and intravoxel incoherent motion imaging for predicting bone metastasis in patients with prostate cancer: a pilot study[J]. Magn Reson Imaging, 2023, 96: 8-16. DOI: 10.1016/j.mri.2022.11.004">10.1016/j.mri.2022.11.004">10.1016/j.mri.2022.11.004.
[3]
LECLER A, DURON L, ZMUDA M, et al. Intravoxel incoherent motion (IVIM) 3 T MRI for orbital lesion characterization[J]. Eur Radiol, 2021, 31(1): 14-23. DOI: 10.1007/s00330-020-07103-1">10.1007/s00330-020-07103-1">10.1007/s00330-020-07103-1.
[4]
YE C, XU D Y, QIN Y B, et al. Estimation of intravoxel incoherent motion parameters using low b-values[J/OL]. PLoS One, 2019, 14(2): e0211911 [2023-01-03]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6364995. DOI: 10.1371/journal.pone.0211911">10.1371/journal.pone.0211911">10.1371/journal.pone.0211911.
[5]
HU Y C, YAN L F, HAN Y, et al. Can the low and high b-value distribution influence the pseudodiffusion parameter derived from IVIM DWI in normal brain?[J/OL]. BMC Med Imaging, 2020, 20(1): 14 [2023-01-03]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7011602. DOI: 10.1186/s12880-020-0419-0">10.1186/s12880-020-0419-0">10.1186/s12880-020-0419-0.
[6]
RYDHÖG A, PASTERNAK O, STÅHLBERG F, et al. Estimation of diffusion, perfusion and fractional volumes using a multi-compartment relaxation-compensated intravoxel incoherent motion (IVIM) signal model[J]. Eur J Radiol Open, 2019, 6: 198-205. DOI: 10.1016/j.ejro.2019.05.007">10.1016/j.ejro.2019.05.007">10.1016/j.ejro.2019.05.007.
[7]
KAANDORP M P T, BARBIERI S, KLAASSEN R, et al. Improved unsupervised physics-informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients[J]. Magn Reson Med, 2021, 86(4): 2250-2265. DOI: 10.1002/mrm.28852">10.1002/mrm.28852">10.1002/mrm.28852.
[8]
VAN DER VELDE N, HASSING H C, BAKKER B J, et al. Improvement of late gadolinium enhancement image quality using a deep learning-based reconstruction algorithm and its influence on myocardial scar quantification[J]. Eur Radiol, 2021, 31(6): 3846-3855. DOI: 10.1007/s00330-020-07461-w">10.1007/s00330-020-07461-w">10.1007/s00330-020-07461-w.
[9]
EICHNER C, PAQUETTE M, MILDNER T, et al. Increased sensitivity and signal-to-noise ratio in diffusion-weighted MRI using multi-echo acquisitions[J/OL]. Neuroimage, 2020, 221: 117172 [2023-01-03]. https://www.sciencedirect.com/science/article/pii/S1053811920306583?via%3Dihub. DOI: 10.1016/j.neuroimage.2020.117172">10.1016/j.neuroimage.2020.117172">10.1016/j.neuroimage.2020.117172.
[10]
DAMEN F C, CAI K J. B1- non-uniformity correction of phased-array coils without measuring coil sensitivity[J]. Magn Reson Imaging, 2018, 51: 20-28. DOI: 10.1016/j.mri.2018.04.009">10.1016/j.mri.2018.04.009">10.1016/j.mri.2018.04.009.
[11]
TAO S, ZHOU X, GRECO E, et al. Edge-enhancing gradient-echo MP2RAGE for clinical epilepsy imaging at 7T[J]. AJNR Am J Neuroradiol, 2023, 44(3): 268-270. DOI: 10.3174/ajnr.A7782">10.3174/ajnr.A7782">10.3174/ajnr.A7782.
[12]
MANDAL P K, GUHA ROY R, KALYANI A. Distribution pattern of closed and extended forms of glutathione in the human brain: MR spectroscopic study[J]. ACS Chem Neurosci, 2023, 14(2): 270-276. DOI: 10.1021/acschemneuro.2c00573">10.1021/acschemneuro.2c00573">10.1021/acschemneuro.2c00573.
[13]
JANG M, JIN S, KANG M, et al. Pattern recognition analysis of directional intravoxel incoherent motion MRI in ischemic rodent brains[J/OL]. NMR Biomed, 2020, 33(5): e4268 [2023-01-09]. https://analyticalsciencejournals.onlinelibrary.wiley.com/doi/10.1002/nbm.4268. DOI: 10.1002/nbm.4268">10.1002/nbm.4268">10.1002/nbm.4268.
[14]
LEE W, CHOI G, LEE J, et al. Registration and quantification network (RQnet) for IVIM-DKI analysis in MRI[J]. Magn Reson Med, 2023, 89(1): 250-261. DOI: 10.1002/mrm.29454">10.1002/mrm.29454">10.1002/mrm.29454.
[15]
IIMA M. Perfusion-driven intravoxel incoherent motion (IVIM) MRI in oncology: applications, challenges, and future trends[J]. Magn Reson Med Sci, 2021, 20(2): 125-138. DOI: 10.2463/mrms.rev.2019-0124">10.2463/mrms.rev.2019-0124">10.2463/mrms.rev.2019-0124.
[16]
BAKKE K M, GRØVIK E, MELTZER S, et al. Comparison of Intravoxel incoherent motion imaging and multiecho dynamic contrast-based MRI in rectal cancer[J]. J Magn Reson Imaging, 2019, 50(4): 1114-1124. DOI: 10.1002/jmri.26740">10.1002/jmri.26740">10.1002/jmri.26740.
[17]
MARAGHECHI B, GACH H M, SETIANEGARA J, et al. Dose uncertainty and resolution of polymer gel dosimetry using an MRI guided radiation therapy system's onboard 0.35 T scanner[J]. Phys Med, 2020, 73: 8-12. DOI: 10.1016/j.ejmp.2020.04.004">10.1016/j.ejmp.2020.04.004">10.1016/j.ejmp.2020.04.004.
[18]
KOOPMANS P J, PFAFFENROT V. Enhanced POCS reconstruction for partial Fourier imaging in multi-echo and time-series acquisitions[J]. Magn Reson Med, 2021, 85(1): 140-151. DOI: 10.1002/mrm.28417">10.1002/mrm.28417">10.1002/mrm.28417.
[19]
MA X F, JIANG G H, LI S M, et al. Aberrant functional connectome in neurologically asymptomatic patients with end-stage renal disease[J/OL]. PLoS One, 2015, 10(3): e0121085 [2023-01-03]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4364738. DOI: 10.1371/journal.pone.0121085">10.1371/journal.pone.0121085">10.1371/journal.pone.0121085.
[20]
WEN Q T, FENG L, ZHOU K, et al. Rapid golden-angle diffusion-weighted propeller MRI for simultaneous assessment of ADC and IVIM[J/OL]. Neuroimage, 2020, 223: 117327 [2023-01-03]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7792631. DOI: 10.1016/j.neuroimage.2020.117327">10.1016/j.neuroimage.2020.117327">10.1016/j.neuroimage.2020.117327.
[21]
STABINSKA J, ZÖLLNER H J, THIEL T, et al. Image Downsampling Expedited Adaptive Least-squares (IDEAL) fitting improves intravoxel incoherent motion (IVIM) analysis in the human kidney[J]. Magn Reson Med, 2023, 89(3): 1055-1067. DOI: 10.1002/mrm.29517">10.1002/mrm.29517">10.1002/mrm.29517.
[22]
OH C, KIM D, CHUNG J Y, et al. A k-space-to-image reconstruction network for MRI using recurrent neural network[J]. Med Phys, 2021, 48(1): 193-203. DOI: 10.1002/mp.14566">10.1002/mp.14566">10.1002/mp.14566.
[23]
LIU B E, ZENG Q Y, HUANG J B, et al. IVIM using convolutional neural networks predicts microvascular invasion in HCC[J]. Eur Radiol, 2022, 32(10): 7185-7195. DOI: 10.1007/s00330-022-08927-9">10.1007/s00330-022-08927-9">10.1007/s00330-022-08927-9.
[24]
VASYLECHKO S D, WARFIELD S K, AFACAN O, et al. Self-supervised IVIM DWI parameter estimation with a physics based forward model[J]. Magn Reson Med, 2022, 87(2): 904-914. DOI: 10.1002/mrm.28989">10.1002/mrm.28989">10.1002/mrm.28989.
[25]
HAHN S, YI J, LEE H J, et al. Image quality and diagnostic performance of accelerated shoulder MRI with deep learning-based reconstruction[J]. AJR Am J Roentgenol, 2022, 218(3): 506-516. DOI: 10.2214/AJR.21.26577">10.2214/AJR.21.26577">10.2214/AJR.21.26577.
[26]
YI D, GRØVIK E, TONG E, et al. MRI pulse sequence integration for deep-learning-based brain metastases segmentation[J]. Med Phys, 2021, 48(10): 6020-6035. DOI: 10.1002/mp.15136">10.1002/mp.15136">10.1002/mp.15136.
[27]
LEE W, KIM B, PARK H. Quantification of intravoxel incoherent motion with optimized b-values using deep neural network[J]. Magn Reson Med, 2021, 86(1): 230-244. DOI: 10.1002/mrm.28708">10.1002/mrm.28708">10.1002/mrm.28708.
[28]
YE C, XU D Y, QIN Y B, et al. Accurate intravoxel incoherent motion parameter estimation using Bayesian fitting and reduced number of low b-values[J]. Med Phys, 2020, 47(9): 4372-4385. DOI: 10.1002/mp.14233">10.1002/mp.14233">10.1002/mp.14233.
[29]
MEEUS E M, NOVAK J, WITHEY S B, et al. Evaluation of intravoxel incoherent motion fitting methods in low-perfused tissue[J]. J Magn Reson Imaging, 2017, 45(5): 1325-1334. DOI: 10.1002/jmri.25411">10.1002/jmri.25411">10.1002/jmri.25411.
[30]
FEDERAU C. Measuring perfusion: intravoxel incoherent motion MR imaging[J]. Magn Reson Imaging Clin N Am, 2021, 29(2): 233-242. DOI: 10.1016/j.mric.2021.01.003">10.1016/j.mric.2021.01.003">10.1016/j.mric.2021.01.003.
[31]
VIENI C, ADES-ARON B, CONTI B, et al. Effect of intravoxel incoherent motion on diffusion parameters in normal brain[J/OL]. Neuroimage, 2020, 204: 116228 [2023-01-03]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6886883. DOI: 10.1016/j.neuroimage.2019.116228">10.1016/j.neuroimage.2019.116228">10.1016/j.neuroimage.2019.116228.
[32]
LEÓN R L, BROWN B P, PERSOHN S A, et al. Intravoxel incoherent motion MR imaging analysis for diagnosis of placenta accrete spectrum disorders: a pilot feasibility study[J]. Magn Reson Imaging, 2021, 80: 26-32. DOI: 10.1016/j.mri.2021.03.007">10.1016/j.mri.2021.03.007">10.1016/j.mri.2021.03.007.
[33]
HAN X P, TAN J, HE Y M. Deep learning algorithm-based MRI image in the diagnosis of diabetic macular edema[J/OL]. Contrast Media Mol Imaging, 2022, 2022: 1-9 [2023-01-03]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8916890. DOI: 10.1155/2022/1035619">10.1155/2022/1035619">10.1155/2022/1035619.

PREV Application of deep learning reconstruction in improving the quality of neuromelanin magnetic resonance image
NEXT A comparative study on the enhancement of high resolution coronal image quality by deep learning reconstruction
  



Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn