Robust Myocardial Perfusion MRI Quantification With DeepFermi

被引:0
|
作者
Brahma, Sherine [1 ]
Kofler, Andreas [1 ]
Zimmermann, Felix F. [1 ]
Schaeffter, Tobias [1 ,2 ,3 ]
Chiribiri, Amedeo [2 ]
Kolbitsch, Christoph [1 ,2 ]
机构
[1] Phys Tech Bundesanstalt, D-38116 Braunschweig, Germany
[2] Kings Coll London, Sch Imaging Sci & Biomed Engn, London, England
[3] Tech Univ Berlin, Dept Med Engn, Einstein Ctr Digital Future, Berlin, Germany
关键词
Myocardium; Training; Optimization; Vectors; Robustness; Signal to noise ratio; Convolutional neural networks; Accuracy; Visualization; Stress; Deep learning; fermi-deconvolution; magnetic resonance imaging; myocardial perfusion quantification; CARDIOVASCULAR MAGNETIC-RESONANCE; EMISSION COMPUTED-TOMOGRAPHY; CORONARY-ARTERY-DISEASE; STATEMENT; IMPACT; NOISE; HEART;
D O I
10.1109/TBME.2024.3485233
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
Stress perfusion cardiac magnetic resonance is an important technique for examining and assessing the blood supply of the myocardium. Currently, the majority of clinical perfusion scans are evaluated based on visual assessment by experienced clinicians. This makes the process subjective, and to this end, quantitative methods have been proposed to offer a more user-independent assessment of perfusion. These methods, however, rely on time-consuming deconvolution analysis and are susceptible to data outliers caused by artifacts due to cardiac or respiratory motion. In our work, we introduce a novel deep-learning method that integrates the commonly used Fermi function with a neural network architecture for fast, accurate, and robust myocardial perfusion quantification. This approach employs the Fermi model to ensure that the perfusion maps are consistent with measured data, while also utilizing a prior based on a 3D convolutional neural network to generalize spatio-temporal information across different patient data. Our network is trained within a self-supervised learning framework, which circumvents the need for ground-truth perfusion labels that are challenging to obtain. Furthermore, we extended this training methodology by adopting a technique that ensures estimations are resistant to data outliers, thereby improving robustness against motion artifacts. Our simulation experiments demonstrated an overall improvement in the accuracy and robustness of perfusion parameter estimation, consistently outperforming traditional deconvolution analysis algorithms across varying Signal-to-Noise Ratio scenarios in the presence of data outliers. For the in vivo studies, our method generated robust perfusion estimates that aligned with clinical diagnoses, while being approximately five times faster than conventional algorithms.
引用
收藏
页码:1031 / 1044
页数:14
相关论文
共 50 条
  • [1] Quantification of Myocardial Perfusion: MRI
    Jerosch-Herold M.
    Shah R.V.
    Current Cardiovascular Imaging Reports, 2012, 5 (3) : 158 - 166
  • [2] Robust estimation of MRI myocardial perfusion parameters
    Aja-Fernandez, Santiago
    Cordero-Grande, Lucilio
    Vegas-Sanchez-Ferrero, Gonzalo
    de Luis-Garcia, Rodrigo
    Alberola-Lopez, Carlos
    2013 35TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC), 2013, : 4382 - 4385
  • [3] Is absolute quantification needed for myocardial perfusion MRI in infarct?
    Vallee, JM
    Lazeyras, F
    Righetti, A
    Didier, D
    RADIOLOGY, 2000, 217 : 240 - 240
  • [4] Quantification of myocardial perfusion by MRI after coronary occlusion
    Vallée, JPM
    Sostman, HD
    MacFall, TR
    DeGrado, TR
    Zhang, J
    Sebbag, L
    Cobb, FR
    Wheeler, T
    Hedlund, LW
    Turkington, TG
    Spritzer, CE
    Coleman, RE
    MAGNETIC RESONANCE IN MEDICINE, 1998, 40 (02) : 287 - 297
  • [5] Quantification of myocardial perfusion with MRI and exogenous contrast agents
    Vallee, JP
    Sostman, HD
    MacFall, JR
    Coleman, RE
    CARDIOLOGY, 1997, 88 (01) : 90 - 105
  • [6] Tracer Kinetic Modeling in Myocardial Perfusion Quantification Using MRI
    Schwab, Felix
    Ingrisch, Michael
    Marcus, Roy
    Bamberg, Fabian
    Hildebrandt, Kristof
    Adrion, Christine
    Gliemi, Christopher
    Nikolaou, Konstantin
    Reiser, Maximilian
    Theisen, Daniel
    MAGNETIC RESONANCE IN MEDICINE, 2015, 73 (03) : 1206 - 1215
  • [7] Physics-informed neural networks for myocardial perfusion MRI quantification
    van Herten R.L.M.
    Chiribiri A.
    Breeuwer M.
    Veta M.
    Scannell C.M.
    Medical Image Analysis, 2022, 78
  • [8] Quantification of myocardial perfusion based on signal intensity of flow sensitized MRI
    Sumeda B Abeykoon
    Janaka Wansapura
    Journal of Cardiovascular Magnetic Resonance, 14 (Suppl 1)
  • [9] Clinical Validation of an Automated Technique for MRI Based Quantification of Myocardial Perfusion
    Tarroni, G.
    Corsi, C.
    Antkowiak, P. F.
    Veronesi, F.
    Kramer, C. M.
    Epstein, F. H.
    Lamberti, C.
    Patel, A. R.
    Mor-Avi, V.
    2011 COMPUTING IN CARDIOLOGY, 2011, 38 : 837 - 840
  • [10] 2133 Effect of the bolus size on the quantification of myocardial perfusion using MRI
    Marko Ivancevic
    Jean-Luc Daire
    Michel Kocher
    Alberto Righetti
    Dominique Didier
    Jean-Paul Vallée
    Journal of Cardiovascular Magnetic Resonance, 10 (Suppl 1)