SPINNED: Simulation-based physics-informed neural network for deconvolution of dynamic susceptibility contrast MRI perfusion data.
DSC MRI
deconvolution
perfusion maps
physics informed neural network
synthetic data
Journal
Magnetic resonance in medicine
ISSN: 1522-2594
Titre abrégé: Magn Reson Med
Pays: United States
ID NLM: 8505245
Informations de publication
Date de publication:
16 Apr 2024
16 Apr 2024
Historique:
revised:
13
03
2024
received:
30
10
2023
accepted:
13
03
2024
medline:
16
4
2024
pubmed:
16
4
2024
entrez:
16
4
2024
Statut:
aheadofprint
Résumé
To propose the simulation-based physics-informed neural network for deconvolution of dynamic susceptibility contrast (DSC) MRI (SPINNED) as an alternative for more robust and accurate deconvolution compared to existing methods. The SPINNED method was developed by generating synthetic tissue residue functions and arterial input functions through mathematical simulations and by using them to create synthetic DSC MRI time series. The SPINNED model was trained using these simulated data to learn the underlying physical relation (deconvolution) between the DSC-MRI time series and the arterial input functions. The accuracy and robustness of the proposed SPINNED method were assessed by comparing it with two common deconvolution methods in DSC MRI data analysis, circulant singular value decomposition, and Volterra singular value decomposition, using both simulation data and real patient data. The proposed SPINNED method was more accurate than the conventional methods across all SNR levels and showed better robustness against noise in both simulation and real patient data. The SPINNED method also showed much faster processing speed than the conventional methods. These results support that the proposed SPINNED method can be a good alternative to the existing methods for resolving the deconvolution problem in DSC MRI. The proposed method does not require any separate ground-truth measurement for training and offers additional benefits of quick processing time and coverage of diverse clinical scenarios. Consequently, it will contribute to more reliable, accurate, and rapid diagnoses in clinical applications compared with the previous methods including those based on supervised learning.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Subventions
Organisme : Korea Medical Device Development Fund
ID : RS-2020-KD000062
Organisme : Hyundai Motor Chung Mong-Koo Foundation Scholarship Program
Organisme : National Research Foundation of Korea
ID : NRF-2020R1A2C2008949
Organisme : National Research Foundation of Korea
ID : RS-2023-00207783
Informations de copyright
© 2024 International Society for Magnetic Resonance in Medicine.
Références
Perkio J, Aronen HJ, Kangasmaki A, et al. Evaluation of four postprocessing methods for determination of cerebral blood volume and mean transit time by dynamic susceptibility contrast imaging. Magn Reson Med. 2002;47:973‐981.
Leiva‐Salinas C, Wintermark M. Imaging of acute ischemic stroke. Neuroimaging Clin N Am. 2010;20:455‐468.
Willats L, Calamante F. The 39 steps: evading error and deciphering the secrets for accurate dynamic susceptibility contrast MRI. NMR Biomed. 2013;26:913‐931.
Wu O, Østergaard L, Sorensen AG. Technical aspects of perfusion‐weighted imaging. Neuroimaging Clin N Am. 2005;15:623‐637.
Sourbron S, Dujardin M, Makkat S, Luypaert R. Pixel‐by‐pixel deconvolution of bolus‐tracking data: optimization and implementation. Phys Med Biol. 2006;52:429‐447.
Ko L, Salluzzi M, Frayne R, Smith M. Reexamining the quantification of perfusion MRI data in the presence of bolus dispersion. J Magn Reson Imaging. 2007;25:639‐643.
Wu O, Østergaard L, Koroshetz WJ, et al. Effects of tracer arrival time on flow estimates in MR perfusion‐weighted imaging. J Magn Reson Imaging. 2003;50:856‐864.
Liu HL, Pu Y, Liu Y, et al. Cerebral blood flow measurement by dynamic contrast MRI using singular value decomposition with an adaptive threshold. J Magn Reson Imaging. 1999;42:167‐172.
Calamante F, Gadian DG, Connelly A. Delay and dispersion effects in dynamic susceptibility contrast MRI: simulations using singular value decomposition. J Magn Reson Imaging. 2000;44:466‐473.
Wu O, Ostergaard L, Weisskoff RM, Benner T, Rosen BR, Sorensen AG. Tracer arrival timing‐insensitive technique for estimating flow in MR perfusion‐weighted imaging using singular value decomposition with a block‐circulant deconvolution matrix. Magn Reson Med. 2003;50:164‐174.
Sourbron S, Luypaert R, Morhard D, Seelos K, Reiser M, Peller M. Deconvolution of bolus‐tracking data: a comparison of discretization methods. Phys Med Biol. 2007;52:6761‐6778.
Zaro‐Weber O, Livne M, Martin SZ, et al. Comparison of the 2 most popular deconvolution techniques for the detection of penumbral flow in acute stroke. Stroke. 2015;46:2795‐2799.
Bell LC, Stokes AM, Quarles CC. Analysis of postprocessing steps for residue function dependent dynamic susceptibility contrast (DSC)‐MRI biomarkers and their clinical impact on glioma grading for both 1.5 and 3T. J Magn Reson Imaging. 2020;51:547‐553.
Schmid N, Bruderer S, Paruzzo F, et al. Deconvolution of 1D NMR spectra: a deep learning‐based approach. J Magn Reson. 2023;347:107357.
Han Y, Kim J, Ye JC. Differentiated backprojection domain deep learning for conebeam artifact removal. IEEE Trans Med Imaging. 2020;39:3571‐3582.
Alaifari R. Ill‐posed problems: from linear to nonlinear and beyond. Harmonic and Applied Analysis: From Radon Transforms to Machine Learning; Springer, 2021:101‐148.
Seo S, Do WJ, Luu HM, Kim KH, Choi SH, Park SH. Artificial neural network for slice encoding for metal artifact correction (SEMAC) MRI. Magn Reson Med. 2020;84:263‐276.
Luu HM, Kim DH, Kim JW, Choi SH, Park SH. qMTNet: accelerated quantitative magnetization transfer imaging with artificial neural networks. Magn Reson Med. 2021;85:298‐308.
Ho KC, Scalzo F, Sarma KV, El‐Saden S, Arnold CW. A temporal deep learning approach for MR perfusion parameter estimation in stroke. Proceedings of 23rd International Conference on Pattern Recognition (ICPR); IEEE; 2016:1315‐1320.
McKinley R, Hung F, Wiest R, Liebeskind DS, Scalzo F. A machine learning approach to perfusion imaging with dynamic susceptibility contrast MR. Front Neurol. 2018;9:717.
Ulas C, Tetteh G, Thrippleton MJ, et al. Direct estimation of pharmacokinetic parameters from DCE‐MRI using deep CNN with forward physical model loss. Proceedings of Medical Image Computing and Computer Assisted Intervention–MICCAI 2018: 21st International Conference, Springer, Granada, Spain; 2018:39‐47.
Meier R, Lux P, Jung S, et al. Neural network–derived perfusion maps for the assessment of lesions in patients with acute ischemic stroke. Radiol Artif Intell. 2019;1:e190019.
Robben D, Suetens P. Perfusion parameter estimation using neural networks and data augmentation. Proceedings of Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries: 4th International Workshop, BrainLes 2018, Held in Conjunction with MICCAI 2018, Springer, Granada, Spain; 2019:439‐446.
Asaduddin M, Roh HG, Kim HJ, Kim EY, Park SH. Perfusion maps acquired from dynamic angiography MRI using deep learning approaches. J Magn Reson Imaging. 2023;57:456‐469.
Asaduddin M, Kim EY, Park SH. A deep learning approach for robust and accurate deconvolution of DSC MRI perfusion calculation. Proceedings of the ISMRM & ISMRT Annual Meeting & Exhibition, 2023, program #2763.
Calamante F, Connelly A, van Osch MJ. Nonlinear ΔR effects in perfusion quantification using bolus‐tracking MRI. Magn Reson Med. 2009;61:486‐492.
Carpenter TK, Armitage PA, Bastin ME, Wardlaw JM. DSC perfusion MRI—quantification and reduction of systematic errors arising in areas of reduced cerebral blood flow. Magn Reson Med. 2006;55:1342‐1349.
Knutsson L, van Westen D, Petersen ET, et al. Absolute quantification of cerebral blood flow: correlation between dynamic susceptibility contrast MRI and model‐free arterial spin labeling. Magn Reson Imaging. 2010;28:1‐7.
Knutsson L, Lindgren E, Ahlgren A, et al. Dynamic susceptibility contrast MRI with a prebolus contrast agent administration design for improved absolute quantification of perfusion. Magn Reson Med. 2014;72:996‐1006.
Asaduddin M, Do WJ, Kim EY, Park SH. Mapping cerebral perfusion from time‐resolved contrast‐enhanced MR angiographic data. Magn Reson Imaging. 2019;61:143‐148.
Østergaard L. Principles of cerebral perfusion imaging by bolus tracking. J Magn Reson Imaging. 2005;22:710‐717.
Kane I, Carpenter T, Chappell F, et al. Comparison of 10 different magnetic resonance perfusion imaging processing methods in acute ischemic stroke: effect on lesion size, proportion of patients with diffusion/perfusion mismatch, clinical scores, and radiologic outcomes. Stroke. 2007;38:3158‐3164.
Calamante F, Christensen S, Desmond PM, Ostergaard L, Davis SM, Connelly A. The physiological significance of the time‐to‐maximum (Tmax) parameter in perfusion MRI. Stroke. 2010;41:1169‐1174.
Meijs M, Christensen S, Lansberg MG, Albers GW, Calamante F. Analysis of perfusion MRI in stroke: to deconvolve, or not to deconvolve. Magn Reson Med. 2016;76:1282‐1290.
Mehndiratta A, Calamante F, MacIntosh BJ, Crane DE, Payne SJ, Chappell MA. Modeling the residue function in DSC‐MRI simulations: analytical approximation to in vivo data. Magn Reson Med. 2014;72:1486‐1491.
Zaharchuk G, Bammer R, Straka M, et al. Improving dynamic susceptibility contrast MRI measurement of quantitative cerebral blood flow using corrections for partial volume and nonlinear contrast relaxivity: a xenon computed tomographic comparative study. J Magn Reson Imaging. 2009;30:743‐752.
Olivot J, Mlynash M, Zaharchuk G, et al. Perfusion MRI (Tmax and MTT) correlation with xenon CT cerebral blood flow in stroke patients. Neurology. 2009;72:1140‐1145.
Shah MK, Shin W, Parikh VS, et al. Quantitative cerebral MR perfusion imaging: preliminary results in stroke. J Magn Reson Imaging. 2010;32:796‐802.
Hess A, Meier R, Kaesmacher J, et al. Synthetic perfusion maps: imaging perfusion deficits in DSC‐MRI with deep learning. Proceedings of Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries: 4th International Workshop, BrainLes 2018, Held in Conjunction with MICCAI, Springer, Granada, Spain; 2018:447‐455.
Andersen IK, Szymkowiak A, Rasmussen CE, et al. Perfusion quantification using gaussian process deconvolution. Magn Reson Med. 2002;48:351‐361.
Wirestam R, Ståhlberg F. Wavelet‐based noise reduction for improved deconvolution of time‐series data in dynamic susceptibility‐contrast MRI. MAGMA. 2005;18:113‐118.
Pizzolato M, Boutelier T, Deriche R. Perfusion deconvolution in DSC‐MRI with dispersion‐compliant bases. Med Image Anal. 2017;36:197‐215.
Chakwizira A, Ahlgren A, Knutsson L, Wirestam R. Non‐parametric deconvolution using Bézier curves for quantification of cerebral perfusion in dynamic susceptibility contrast MRI. MAGMA. 2022;35:791‐804.
Molnar M, Reardon KP, Osborne C, Milić I. Spectral deconvolution with deep learning: removing the effects of spectral PSF broadening. Front Astron Space Sci. 2020;7:29.
Østergaard L, Chesler DA, Weisskoff RM, Sorensen AG, Rosen BR. Modeling cerebral blood flow and flow heterogeneity from magnetic resonance residue data. J Cereb Blood Flow Metab. 1999;19:690‐699.