*Result*: Unsupervised 1D CNN -bidirectional long short-term memory model with multi-head attention for generating intravoxel incoherent motion maps.
Title:
Unsupervised 1D CNN -bidirectional long short-term memory model with multi-head attention for generating intravoxel incoherent motion maps.
Authors:
Li ZY; Institute of Medical Device and Imaging, College of Medicine, National Taiwan University, Taipei City, Taiwan.; Department of Radiology, The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, Zhejiang, China., Huang HM; Institute of Medical Device and Imaging, College of Medicine, National Taiwan University, Taipei City, Taiwan.; Program for Precision Health and Intelligent Medicine, Graduate School of Advanced Technology, National Taiwan University, Taipei City, Taiwan.
Source:
Medical physics [Med Phys] 2026 Mar; Vol. 53 (3), pp. e70396.
Publication Type:
Journal Article
Language:
English
Journal Info:
Publisher: John Wiley and Sons, Inc Country of Publication: United States NLM ID: 0425746 Publication Model: Print Cited Medium: Internet ISSN: 2473-4209 (Electronic) Linking ISSN: 00942405 NLM ISO Abbreviation: Med Phys Subsets: MEDLINE
Imprint Name(s):
Publication: 2017- : Hoboken, NJ : John Wiley and Sons, Inc.
Original Publication: Lancaster, Pa., Published for the American Assn. of Physicists in Medicine by the American Institute of Physics.
Original Publication: Lancaster, Pa., Published for the American Assn. of Physicists in Medicine by the American Institute of Physics.
MeSH Terms:
References:
De Luca A, Bertoldo A, Froeling M. Effects of perfusion on DTI and DKI estimates in the skeletal muscle. Magn Reson Med. 2017;78(1):233‐246. doi:10.1002/mrm.26373.
Le Bihan D, Breton E, Lallemand D, et al. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology. 1988;168(2):497‐505. doi:10.1148/radiology.168.2.3393671.
Einstein A. Investigations on the Theory of the Brownian Movement. Courier Corporation; 1956:119.
Le Bihan D, Breton E, Lallemand D, et al. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology. 1986;161(2):401‐407. doi:10.1148/radiology.161.2.3763909.
Liu C, Liang C, Liu Z, Zhang S, Huang B. Intravoxel incoherent motion (IVIM) in evaluation of breast lesions: comparison with conventional DWI. Eur J Radiol. 2013;82(12):e782‐789. doi:10.1016/j.ejrad.2013.08.006.
Chiaradia M, Baranes L, Van Nhieu JT, et al. Intravoxel incoherent motion (IVIM) MR imaging of colorectal liver metastases: are we only looking at tumor necrosis?. J Magn Reson Imaging. 2014;39(2):317‐325. doi:10.1002/jmri.24172.
Peng J, Zheng J, Yang C, et al. Intravoxel incoherent motion diffusion‐weighted imaging to differentiate hepatocellular carcinoma from intrahepatic cholangiocarcinoma. Sci Rep. 2020;10(1):7717. doi:10.1038/s41598‐020‐64804‐9.
Federau C, O'Brien K, Meuli R, Hagmann P, Maeder P. Measuring brain perfusion with intravoxel incoherent motion (IVIM): initial clinical experience. J Magn Reson Imaging. 2014;39(3):624‐632. doi:10.1002/jmri.24195.
Spinner GR, Federau C, Kozerke S. Bayesian inference using hierarchical and spatial priors for intravoxel incoherent motion MR imaging in the brain: analysis of cancer and acute stroke. Med Image Anal. 2021;73:102144. doi:10.1016/j.media.2021.102144.
Neil JJ, Bosch CS, Ackerman JJ. An evaluation of the sensitivity of the intravoxel incoherent motion (IVIM) method of blood flow measurement to changes in cerebral blood flow. Magn Reson Med. 1994;32(1):60‐65. doi:10.1002/mrm.1910320109.
Slator PJ, Hutter J, McCabe L, et al. Placenta microstructure and microcirculation imaging with diffusion MRI. Magn Reson Med. 2018;80(2):756‐766. doi:10.1002/mrm.27036.
Pekar J, Moonen CT, van Zijl PC. On the precision of diffusion/perfusion imaging by gradient sensitization. Magn Reson Med. 1992;23(1):122‐129. doi:10.1002/mrm.1910230113.
Luciani A, Vignaud A, Cavet M, et al. Liver cirrhosis: intravoxel incoherent motion MR imaging–pilot study. Radiology. 2008;249(3):891‐899. doi:10.1148/radiol.2493080080.
Cho GY, Moy L, Zhang JL, et al. Comparison of fitting methods and b‐value sampling strategies for intravoxel incoherent motion in breast cancer. Magn Reson Med. 2015;74(4):1077‐1085. doi:10.1002/mrm.25484.
Ogura A, Hayakawa K, Miyati T, Maeda F. Imaging parameter effects in apparent diffusion coefficient determination of magnetic resonance imaging. Eur J Radiol. 2011;77(1):185‐188. doi:10.1016/j.ejrad.2009.06.031.
Branch MA, Coleman TF, Li YY. A subspace, interior, and conjugate gradient method for large‐scale bound‐constrained minimization problems. SIAM J Sci Comput. 1999;21(1):1‐23. doi:10.1137/S1064827595289108.
Byrd RH, Schnabel RB, Shultz GA. Approximate solution of the trust region problem by minimization over two‐dimensional subspaces. Math Program. 1988;40(1):247‐263. doi:10.1007/BF01580735.
Orton MR, Collins DJ, Koh DM, Leach MO. Improved intravoxel incoherent motion analysis of diffusion weighted imaging by data driven Bayesian modeling. Magn Reson Med. 2014;71(1):411‐420. doi:10.1002/mrm.24649.
Barbieri S, Donati OF, Froehlich JM, Thoeny HC. Impact of the calculation algorithm on biexponential fitting of diffusion‐weighted MRI in upper abdominal organs. Magn Reson Med. 2016;75(5):2175‐2184. doi:10.1002/mrm.25765.
Marin JM, Pudlo P, Robert CP, Ryder RJ. Approximate Bayesian computational methods. Stat Comput. 2012;22(6):1167‐1180. doi:10.1007/s11222‐011‐9288‐2.
Mastropietro A, Procissi D, Scalco E, Rizzo G, Bertolino N. A supervised deep neural network approach with standardized targets for enhanced accuracy of IVIM parameter estimation from multi‐SNR images. NMR Biomed. 2022;35(10):e4774. doi:10.1002/nbm.4774.
Lee W, Kim B, Park H. Quantification of intravoxel incoherent motion with optimized b‐values using deep neural network. Magn Reson Med. 2021;86(1):230‐244. doi:10.1002/mrm.28708.
Barbieri S, Gurney‐Champion OJ, Klaassen R, Thoeny HC. Deep learning how to fit an intravoxel incoherent motion model to diffusion‐weighted MRI. Magn Reson Med. 2020;83(1):312‐321. doi:10.1002/mrm.27910.
Kaandorp MPT, Barbieri S, Klaassen R, et al. Improved unsupervised physics‐informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients. Magn Reson Med. 2021;86(4):2250‐2265. doi:10.1002/mrm.28852.
Summers C, Dinneen MJ, Nondeterminism and instability in neural network optimization. International Conference on Machine Learning. 2021:9913‐9922.
Ozaki Y, Yano M, Onishi M. Effective hyperparameter optimization using Nelder‐Mead method in deep learning. IPSJ Trans Comput Vis Appl. 2017;9(1):1‐12. doi:10.1186/s41074‐017‐0030‐7.
Isaksson LJ, Repetto M, Summers PE, et al. High‐performance prediction models for prostate cancer radiomics. Inf Med Unlock. 2023;37:101161. doi:10.1016/j.imu.2023.101161.
Hochreiter S, Schmidhuber J. Long short‐term memory. Neural Comput. 1997;9(8):1735‐1780. doi:10.1162/neco.1997.9.8.1735.
Vaswani A, Shaeeer N, Parmar N, et al. Attention is all you need. Advances in Neural Information Processing Systems. 2017:5998‐6008.
Albawi S, Mohammed TA, Al‐Zawi S. Understanding of a convolutional neural network. Proc Int Conf Eng Technol. 2017:1‐6.
Luong M‐T, Pham H, Manning CD. Effective approaches to attention‐based neural machine translation. arXiv. 2015:150804025. doi:10.48550/arXiv.1508.04025.
Aja‐Fernandez S, Alberola‐Lopez C, Westin CF. Noise and signal estimation in magnitude MRI and Rician distributed images: a LMMSE approach. IEEE Trans Image Process. 2008;17(8):1383‐1398. doi:10.1109/TIP.2008.925382.
Manjón JV, Coupé P, Concha L, Buades A, Collins DL, Robles M. Diffusion weighted image denoising using overcomplete local PCA. PLoS One. 2013;8:e73021.
Koay CG, Basser PJ. Analytically exact correction scheme for signal extraction from noisy magnitude MR signals. J Magn Reson. 2006;179(2):317‐322. doi:10.1016/j.jmr.2006.01.016.
Coupe P, Yger P, Prima S, et al. An optimized blockwise nonlocal means denoising filter for 3‐D magnetic resonance images. IEEE Trans Med Imaging. 2008;27(4):425‐441. doi:10.1109/TMI.2007.906087.
Ashby SF, Holst MJ, Manteuffel TA, Saylor PE. The role of the inner product in stopping criteria for conjugate gradient iterations. BIT Numer Math. 2001;41(1):26‐52. doi:10.1023/A:1021961616621.
Kingma DP, Adam BaJ. A method for stochastic optimization. arXiv. 2014:1412.6980. doi:10.48550/arXiv.1412.6980.
Sigmund EE, Vivier PH, Sui D, et al. Intravoxel incoherent motion and diffusion‐tensor imaging in renal tissue under hydration and furosemide flow challenges. Radiology. 2012;263(3):758‐769. doi:10.1148/radiol.12111327.
Kang KM, Lee JM, Yoon JH, et al. Intravoxel incoherent motion diffusion‐weighted MR imaging for characterization of focal pancreatic lesions. Radiology. 2014;270(2):444‐453. doi:10.1148/radiol.13122712.
Taimouri V, Afacan O, Perez‐Rossello JM, et al. Spatially constrained incoherent motion method improves diffusion‐weighted MRI signal decay analysis in the liver and spleen. Med Phys. 2015;42(4):1895‐1903. doi:10.1118/1.4915495.
Wetscherek A, Stieltjes B, Laun FB. Flow‐compensated intravoxel incoherent motion diffusion imaging. Magn Reson Med. 2015;74(2):410‐419. doi:10.1002/mrm.25410.
Lu P‐X, Huang H, Yuan J, et al. Decreases in molecular diffusion, perfusion fraction and perfusion‐related diffusion in fibrotic livers: a prospective clinical intravoxel incoherent motion MR imaging study. PLoS One. 2014;9(12):e113846. doi:10.1371/journal.pone.0113846.
van Baalen S, Leemans A, Dik P, et al. Intravoxel incoherent motion modeling in the kidneys: comparison of mono‐, bi‐, and triexponential fit. J Magn Reson Imaging. 2017;46(1):228‐239. doi:10.1002/jmri.25519.
Gurney‐Champion OJ, Klaassen R, Froeling M, et al. Comparison of six fit algorithms for the intra‐voxel incoherent motion model of diffusion‐weighted magnetic resonance imaging data of pancreatic cancer patients. PLoS One. 2018;13(4):e0194590. doi:10.1371/journal.pone.0194590.
Klaassen R, Gurney‐Champion OJ, Engelbrecht MRW, et al. Evaluation of six diffusion‐weighted MRI models for assessing effects of neoadjuvant chemoradiation in pancreatic cancer patients. Int J Radiat Oncol Biol Phys. 2018;102(4):1052‐1062. doi:10.1016/j.ijrobp.2018.04.064.
Huang HM. An unsupervised convolutional neural network method for estimation of intravoxel incoherent motion parameters. Phys Med Biol. 2022;67(21):215018. doi:10.1088/1361‐6560/ac9a1f.
Rydhog AS, van Osch MJ, Lindgren E, et al. Intravoxel incoherent motion (IVIM) imaging at different magnetic field strengths: what is feasible?. Magn Reson Imaging. 2014;32(10):1247‐1258. doi:10.1016/j.mri.2014.07.013.
Wu WC, Chen YF, Tseng HM, Yang SC, My PC. Caveat of measuring perfusion indexes using intravoxel incoherent motion magnetic resonance imaging in the human brain. Eur Radiol. 2015;25(8):2485‐2492. doi:10.1007/s00330‐015‐3655‐x.
Vieni C, Ades‐Aron B, Conti B, et al. Effect of intravoxel incoherent motion on diffusion parameters in normal brain. Neuroimage. 2020;204:116228. doi:10.1016/j.neuroimage.2019.116228.
Wang H, Li T, Zhuang Z, et al. Early stopping for deep image prior. arXiv. 2021:211206074.
Shi ZL, Mettes P, Maji S, Snoek CGM. On measuring and controlling the spectral bias of the deep image prior. Int J Comput Vision. 2022;130(4):885‐908. doi:10.1007/s11263‐021‐01572‐7.
Le Bihan D, Breton E, Lallemand D, et al. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology. 1988;168(2):497‐505. doi:10.1148/radiology.168.2.3393671.
Einstein A. Investigations on the Theory of the Brownian Movement. Courier Corporation; 1956:119.
Le Bihan D, Breton E, Lallemand D, et al. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology. 1986;161(2):401‐407. doi:10.1148/radiology.161.2.3763909.
Liu C, Liang C, Liu Z, Zhang S, Huang B. Intravoxel incoherent motion (IVIM) in evaluation of breast lesions: comparison with conventional DWI. Eur J Radiol. 2013;82(12):e782‐789. doi:10.1016/j.ejrad.2013.08.006.
Chiaradia M, Baranes L, Van Nhieu JT, et al. Intravoxel incoherent motion (IVIM) MR imaging of colorectal liver metastases: are we only looking at tumor necrosis?. J Magn Reson Imaging. 2014;39(2):317‐325. doi:10.1002/jmri.24172.
Peng J, Zheng J, Yang C, et al. Intravoxel incoherent motion diffusion‐weighted imaging to differentiate hepatocellular carcinoma from intrahepatic cholangiocarcinoma. Sci Rep. 2020;10(1):7717. doi:10.1038/s41598‐020‐64804‐9.
Federau C, O'Brien K, Meuli R, Hagmann P, Maeder P. Measuring brain perfusion with intravoxel incoherent motion (IVIM): initial clinical experience. J Magn Reson Imaging. 2014;39(3):624‐632. doi:10.1002/jmri.24195.
Spinner GR, Federau C, Kozerke S. Bayesian inference using hierarchical and spatial priors for intravoxel incoherent motion MR imaging in the brain: analysis of cancer and acute stroke. Med Image Anal. 2021;73:102144. doi:10.1016/j.media.2021.102144.
Neil JJ, Bosch CS, Ackerman JJ. An evaluation of the sensitivity of the intravoxel incoherent motion (IVIM) method of blood flow measurement to changes in cerebral blood flow. Magn Reson Med. 1994;32(1):60‐65. doi:10.1002/mrm.1910320109.
Slator PJ, Hutter J, McCabe L, et al. Placenta microstructure and microcirculation imaging with diffusion MRI. Magn Reson Med. 2018;80(2):756‐766. doi:10.1002/mrm.27036.
Pekar J, Moonen CT, van Zijl PC. On the precision of diffusion/perfusion imaging by gradient sensitization. Magn Reson Med. 1992;23(1):122‐129. doi:10.1002/mrm.1910230113.
Luciani A, Vignaud A, Cavet M, et al. Liver cirrhosis: intravoxel incoherent motion MR imaging–pilot study. Radiology. 2008;249(3):891‐899. doi:10.1148/radiol.2493080080.
Cho GY, Moy L, Zhang JL, et al. Comparison of fitting methods and b‐value sampling strategies for intravoxel incoherent motion in breast cancer. Magn Reson Med. 2015;74(4):1077‐1085. doi:10.1002/mrm.25484.
Ogura A, Hayakawa K, Miyati T, Maeda F. Imaging parameter effects in apparent diffusion coefficient determination of magnetic resonance imaging. Eur J Radiol. 2011;77(1):185‐188. doi:10.1016/j.ejrad.2009.06.031.
Branch MA, Coleman TF, Li YY. A subspace, interior, and conjugate gradient method for large‐scale bound‐constrained minimization problems. SIAM J Sci Comput. 1999;21(1):1‐23. doi:10.1137/S1064827595289108.
Byrd RH, Schnabel RB, Shultz GA. Approximate solution of the trust region problem by minimization over two‐dimensional subspaces. Math Program. 1988;40(1):247‐263. doi:10.1007/BF01580735.
Orton MR, Collins DJ, Koh DM, Leach MO. Improved intravoxel incoherent motion analysis of diffusion weighted imaging by data driven Bayesian modeling. Magn Reson Med. 2014;71(1):411‐420. doi:10.1002/mrm.24649.
Barbieri S, Donati OF, Froehlich JM, Thoeny HC. Impact of the calculation algorithm on biexponential fitting of diffusion‐weighted MRI in upper abdominal organs. Magn Reson Med. 2016;75(5):2175‐2184. doi:10.1002/mrm.25765.
Marin JM, Pudlo P, Robert CP, Ryder RJ. Approximate Bayesian computational methods. Stat Comput. 2012;22(6):1167‐1180. doi:10.1007/s11222‐011‐9288‐2.
Mastropietro A, Procissi D, Scalco E, Rizzo G, Bertolino N. A supervised deep neural network approach with standardized targets for enhanced accuracy of IVIM parameter estimation from multi‐SNR images. NMR Biomed. 2022;35(10):e4774. doi:10.1002/nbm.4774.
Lee W, Kim B, Park H. Quantification of intravoxel incoherent motion with optimized b‐values using deep neural network. Magn Reson Med. 2021;86(1):230‐244. doi:10.1002/mrm.28708.
Barbieri S, Gurney‐Champion OJ, Klaassen R, Thoeny HC. Deep learning how to fit an intravoxel incoherent motion model to diffusion‐weighted MRI. Magn Reson Med. 2020;83(1):312‐321. doi:10.1002/mrm.27910.
Kaandorp MPT, Barbieri S, Klaassen R, et al. Improved unsupervised physics‐informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients. Magn Reson Med. 2021;86(4):2250‐2265. doi:10.1002/mrm.28852.
Summers C, Dinneen MJ, Nondeterminism and instability in neural network optimization. International Conference on Machine Learning. 2021:9913‐9922.
Ozaki Y, Yano M, Onishi M. Effective hyperparameter optimization using Nelder‐Mead method in deep learning. IPSJ Trans Comput Vis Appl. 2017;9(1):1‐12. doi:10.1186/s41074‐017‐0030‐7.
Isaksson LJ, Repetto M, Summers PE, et al. High‐performance prediction models for prostate cancer radiomics. Inf Med Unlock. 2023;37:101161. doi:10.1016/j.imu.2023.101161.
Hochreiter S, Schmidhuber J. Long short‐term memory. Neural Comput. 1997;9(8):1735‐1780. doi:10.1162/neco.1997.9.8.1735.
Vaswani A, Shaeeer N, Parmar N, et al. Attention is all you need. Advances in Neural Information Processing Systems. 2017:5998‐6008.
Albawi S, Mohammed TA, Al‐Zawi S. Understanding of a convolutional neural network. Proc Int Conf Eng Technol. 2017:1‐6.
Luong M‐T, Pham H, Manning CD. Effective approaches to attention‐based neural machine translation. arXiv. 2015:150804025. doi:10.48550/arXiv.1508.04025.
Aja‐Fernandez S, Alberola‐Lopez C, Westin CF. Noise and signal estimation in magnitude MRI and Rician distributed images: a LMMSE approach. IEEE Trans Image Process. 2008;17(8):1383‐1398. doi:10.1109/TIP.2008.925382.
Manjón JV, Coupé P, Concha L, Buades A, Collins DL, Robles M. Diffusion weighted image denoising using overcomplete local PCA. PLoS One. 2013;8:e73021.
Koay CG, Basser PJ. Analytically exact correction scheme for signal extraction from noisy magnitude MR signals. J Magn Reson. 2006;179(2):317‐322. doi:10.1016/j.jmr.2006.01.016.
Coupe P, Yger P, Prima S, et al. An optimized blockwise nonlocal means denoising filter for 3‐D magnetic resonance images. IEEE Trans Med Imaging. 2008;27(4):425‐441. doi:10.1109/TMI.2007.906087.
Ashby SF, Holst MJ, Manteuffel TA, Saylor PE. The role of the inner product in stopping criteria for conjugate gradient iterations. BIT Numer Math. 2001;41(1):26‐52. doi:10.1023/A:1021961616621.
Kingma DP, Adam BaJ. A method for stochastic optimization. arXiv. 2014:1412.6980. doi:10.48550/arXiv.1412.6980.
Sigmund EE, Vivier PH, Sui D, et al. Intravoxel incoherent motion and diffusion‐tensor imaging in renal tissue under hydration and furosemide flow challenges. Radiology. 2012;263(3):758‐769. doi:10.1148/radiol.12111327.
Kang KM, Lee JM, Yoon JH, et al. Intravoxel incoherent motion diffusion‐weighted MR imaging for characterization of focal pancreatic lesions. Radiology. 2014;270(2):444‐453. doi:10.1148/radiol.13122712.
Taimouri V, Afacan O, Perez‐Rossello JM, et al. Spatially constrained incoherent motion method improves diffusion‐weighted MRI signal decay analysis in the liver and spleen. Med Phys. 2015;42(4):1895‐1903. doi:10.1118/1.4915495.
Wetscherek A, Stieltjes B, Laun FB. Flow‐compensated intravoxel incoherent motion diffusion imaging. Magn Reson Med. 2015;74(2):410‐419. doi:10.1002/mrm.25410.
Lu P‐X, Huang H, Yuan J, et al. Decreases in molecular diffusion, perfusion fraction and perfusion‐related diffusion in fibrotic livers: a prospective clinical intravoxel incoherent motion MR imaging study. PLoS One. 2014;9(12):e113846. doi:10.1371/journal.pone.0113846.
van Baalen S, Leemans A, Dik P, et al. Intravoxel incoherent motion modeling in the kidneys: comparison of mono‐, bi‐, and triexponential fit. J Magn Reson Imaging. 2017;46(1):228‐239. doi:10.1002/jmri.25519.
Gurney‐Champion OJ, Klaassen R, Froeling M, et al. Comparison of six fit algorithms for the intra‐voxel incoherent motion model of diffusion‐weighted magnetic resonance imaging data of pancreatic cancer patients. PLoS One. 2018;13(4):e0194590. doi:10.1371/journal.pone.0194590.
Klaassen R, Gurney‐Champion OJ, Engelbrecht MRW, et al. Evaluation of six diffusion‐weighted MRI models for assessing effects of neoadjuvant chemoradiation in pancreatic cancer patients. Int J Radiat Oncol Biol Phys. 2018;102(4):1052‐1062. doi:10.1016/j.ijrobp.2018.04.064.
Huang HM. An unsupervised convolutional neural network method for estimation of intravoxel incoherent motion parameters. Phys Med Biol. 2022;67(21):215018. doi:10.1088/1361‐6560/ac9a1f.
Rydhog AS, van Osch MJ, Lindgren E, et al. Intravoxel incoherent motion (IVIM) imaging at different magnetic field strengths: what is feasible?. Magn Reson Imaging. 2014;32(10):1247‐1258. doi:10.1016/j.mri.2014.07.013.
Wu WC, Chen YF, Tseng HM, Yang SC, My PC. Caveat of measuring perfusion indexes using intravoxel incoherent motion magnetic resonance imaging in the human brain. Eur Radiol. 2015;25(8):2485‐2492. doi:10.1007/s00330‐015‐3655‐x.
Vieni C, Ades‐Aron B, Conti B, et al. Effect of intravoxel incoherent motion on diffusion parameters in normal brain. Neuroimage. 2020;204:116228. doi:10.1016/j.neuroimage.2019.116228.
Wang H, Li T, Zhuang Z, et al. Early stopping for deep image prior. arXiv. 2021:211206074.
Shi ZL, Mettes P, Maji S, Snoek CGM. On measuring and controlling the spectral bias of the deep image prior. Int J Comput Vision. 2022;130(4):885‐908. doi:10.1007/s11263‐021‐01572‐7.
Grant Information:
NSTC114-2221-E-002-093 National Science and Technology Council, Taiwan
Contributed Indexing:
Keywords: convolutional neural network; diffusion MRI; intravoxel incoherent motion imaging; unsupervised learning
Entry Date(s):
Date Created: 20260319 Date Completed: 20260319 Latest Revision: 20260319
Update Code:
20260320
DOI:
10.1002/mp.70396
PMID:
41855015
Database:
MEDLINE
*Further Information*
*Background: Intravoxel incoherent motion (IVIM) imaging, a diffusion magnetic resonance imaging technique, is commonly used to quantify tissue perfusion and diffusion. Traditional pixel-by-pixel fitting methods, however, often suffer from high noise, causing unreliable parameter estimates.
Purpose and Methods: To address this issue, a novel unsupervised learning-based framework combining a one-dimensional (1D) convolutional neural network (CNN) with a bidirectional long short-term memory (BiLSTM) network and a multi-head attention mechanism (MHAM) was proposed. Several techniques were proposed to reduce the effect of random weight initialization, noisy input data, and overfitting/underfitting on the estimation of IVIM parameters. The performance of the proposed method was evaluated using both simulated and experimental data, and the results were compared with those obtained using the deep neural network (DNN) method and the Bayesian-Markov random fields (MRF) method.
Results: Simulation results showed that the proposed method achieved lower root mean square error values than the other two methods, indicating more reliable IVIM parameter estimates. The only exception was at a signal-to-noise ratio of 100, where it performed similarly to the Bayesian-MRF method. For the abdominal datasets, the proposed method yielded IVIM parameter estimates that closely matched the literature-reported values and avoided the overestimation of pseudo-diffusion coefficients (D<sup>*</sup>) observed in the other two methods. For the brain dataset, the perfusion fractions and diffusion coefficients obtained from all three methods were consistent with the literature-reported ranges; however, only the DNN method tended to overestimate D<sup>*</sup>.
Conclusions: These findings suggest that the proposed CNN-BiLSTM-MHAM model is a promising approach for IVIM parameter estimation.
(© 2026 American Association of Physicists in Medicine.)*