Incorporation of a spectral model in a convolutional neural network for accelerated spectral fitting

• Type of Material: Article
• Authors: Saumya S. Gurbani, Emory University; Sulaiman Sheriff, University of Miami; Andrew A. Maudsley, University of Miami; Hyunsuk Shim, Emory University; Lee Cooper, Emory University
• Language: English
• Date: 2019-05-01
• Publisher: Wiley
• Publication Version: Author Accepted Manuscript (After Peer Review)
• Copyright Statement: © 2019 John Wiley & Sons, Inc. All rights reserved.
• Final Published Version (URL): http://dx.doi.org/10.1002/mrm.27641
• Title of Journal or Parent Work: Magnetic Resonance in Medicine
• Volume: 81
• Issue: 5
• Start Page: 3346
• End Page: 3357
• Grant/Funding Information: This work was supported by National Institutes of Health grants: U01 EB028145, R01 CA214557, R01 EB016064, and F30 CA206291.
• Keywords: spectral analysis; spectroscopic imaging; Gliomas; deep learning; machine learning; MR spectroscopy; Life Sciences & Biomedicine; brain; Automated quantitation; Quality; Metabolites; Radiation therapy; Quantification; Radiology, Nuclear Medicine & Medical Imaging; Magnetic resonance spectroscopy; Science & Technology; MRSI
• Research Categories: Health Sciences, Oncology; Health Sciences, Radiology; Engineering, Biomedical; Biology, Neuroscience

Abstract

Purpose: MRSI has shown great promise in the detection and monitoring of neurologic pathologies such as tumor. A necessary component of data processing includes the quantitation of each metabolite, typically done through fitting a model of the spectrum to the data. For high-resolution volumetric MRSI of the brain, which may have ~10,000 spectra, significant processing time is required for spectral analysis and generation of metabolite maps. Methods: A novel unsupervised deep learning architecture that combines a convolutional neural network with a priori models of the spectrum is presented. This architecture, a convolutional encoder–model decoder (CEMD), combines the strengths of adaptive and unbiased convolutional networks with models of magnetic resonance and is readily interpretable. Results: The CEMD architecture performs accurate spectral fitting for volumetric MRSI in patients with glioblastoma, provides whole-brain fitting in 1 min on a standard computer, and handles a variety of spectral artifacts. Conclusion: A new architecture combining physics domain knowledge with convolutional neural networks has been developed and is able to perform rapid spectral fitting of whole-brain data. Rapid processing is a critical step toward routine clinical practice.