Deep Learning-Based Denoising of CEST MR Data: A Feasibility Study on Applying Synthetic Phantoms in Medical Imaging
-
Published:2023-10-27
Issue:21
Volume:13
Page:3326
-
ISSN:2075-4418
-
Container-title:Diagnostics
-
language:en
-
Short-container-title:Diagnostics
Author:
Radke Karl Ludger1ORCID, Kamp Benedikt1ORCID, Adriaenssens Vibhu1, Stabinska Julia23, Gallinnis Patrik1, Wittsack Hans-Jörg1, Antoch Gerald1, Müller-Lutz Anja1
Affiliation:
1. Department of Diagnostic and Interventional Radiology, Medical Faculty, University Dusseldorf, 40225 Dusseldorf, Germany 2. F.M. Kirby Research Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, MD 21205, USA 3. Division of MR Research, The Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
Abstract
Chemical Exchange Saturation Transfer (CEST) magnetic resonance imaging (MRI) provides a novel method for analyzing biomolecule concentrations in tissues without exogenous contrast agents. Despite its potential, achieving a high signal-to-noise ratio (SNR) is imperative for detecting small CEST effects. Traditional metrics such as Magnetization Transfer Ratio Asymmetry (MTRasym) and Lorentzian analyses are vulnerable to image noise, hampering their precision in quantitative concentration estimations. Recent noise-reduction algorithms like principal component analysis (PCA), nonlocal mean filtering (NLM), and block matching combined with 3D filtering (BM3D) have shown promise, as there is a burgeoning interest in the utilization of neural networks (NNs), particularly autoencoders, for imaging denoising. This study uses the Bloch–McConnell equations, which allow for the synthetic generation of CEST images and explores NNs efficacy in denoising these images. Using synthetically generated phantoms, autoencoders were created, and their performance was compared with traditional denoising methods using various datasets. The results underscored the superior performance of NNs, notably the ResUNet architectures, in noise identification and abatement compared to analytical approaches across a wide noise gamut. This superiority was particularly pronounced at elevated noise intensities in the in vitro data. Notably, the neural architectures significantly improved the PSNR values, achieving up to 35.0, while some traditional methods struggled, especially in low-noise reduction scenarios. However, the application to the in vivo data presented challenges due to varying noise profiles. This study accentuates the potential of NNs as robust denoising tools, but their translation to clinical settings warrants further investigation.
Funder
Jürgen Manchot Stiftung
Subject
Clinical Biochemistry
Reference54 articles.
1. Vinogradov, E., Keupp, J., Dimitrov, I.E., Seiler, S., and Pedrosa, I. (2023). CEST-MRI for body oncologic imaging: Are we there yet?. NMR Biomed., 36. 2. Lingl, J.P., Wunderlich, A., Goerke, S., Paech, D., Ladd, M.E., Liebig, P., Pala, A., Kim, S.Y., Braun, M., and Schmitz, B.L. (2022). The Value of APTw CEST MRI in Routine Clinical Assessment of Human Brain Tumor Patients at 3T. Diagnostics, 12. 3. Chemical exchange saturation transfer (CEST): What is in a name and what isn’t?;Yadav;Magn. Reson. Med.,2011 4. High-sensitivity CEST mapping using a spatiotemporal correlation-enhanced method;Chen;Magn. Reson. Med.,2020 5. Radke, K.L., Wilms, L.M., Frenken, M., Stabinska, J., Knet, M., Kamp, B., Thiel, T.A., Filler, T.J., Nebelung, S., and Antoch, G. (2022). Lorentzian-Corrected Apparent Exchange-Dependent Relaxation (LAREX) Ω-Plot Analysis-An Adaptation for qCEST in a Multi-Pool System: Comprehensive In Silico, In Situ, and In Vivo Studies. Int. J. Mol. Sci., 23.
|
|