News | Artificial Intelligence | July 17, 2018

Artificial Intelligence Provides Faster, Clearer MRI Scans

Machine learning software produces better signal and less noise than conventional MRI techniques

Artificial Intelligence Provides Faster, Clearer MRI Scans

A new artificial-intelligence-based approach to image reconstruction, called AUTOMAP, yields higher quality images from less data, reducing radiation doses for CT and PET and shortening scan times for MRI. Shown here are MR images reconstructed from the same data with conventional approaches, at left, and AUTOMAP, at right. Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital

July 17, 2018 — A research team with funding from the National Institute for Biomedical Imaging and Bioengineering (NIBIB) has developed an advanced computing technique for rapidly and cost effectively improving the quality of biomedical imaging. The technology, called AUTOMAP, uses machine learning and software, referred to as neural networks — inspired by the brain’s ability to process information and perceive or make choices. AUTOMAP finds the best computational strategies to produce clear, accurate images for various types of medical scans.

In their study in the March 21, 2018, issue of Nature, the researchers from Massachusetts General Hospital (MGH) Martinos Center for Biomedical Imaging and Harvard University found that the AUTOMAP system could produce brain magnetic resonance imaging (MRI) images with better signal and less noise than conventional MRI techniques. Achieving a good signal-to-noise ratio is a key factor in generating a quality MRI scan.

MRI uses a magnetic field and radio waves to create detailed images of tissues inside the body. Noise from either electronic sources or from tissues in the body are detrimental to image quality, so imagers look for ways to lessen its effect.

“The signal-to-noise ratio improvements we gain from this artificial intelligence-based method directly accelerates image acquisition on low-field MRI,” said lead author Bo Zhu, Ph.D., postdoctoral research fellow in radiology at Harvard Medical School and in physics at the MGH Martinos Center. NIBIB has supported Zhu’s postdoctoral research on this project. He added that the AUTOMAP neural network will be compatible with novel image acquisition strategies and unconventional hardware designs.

AUTOMAP churns through — and learns from — data from existing images and applies mathematical approaches in reconstructing new ones. The team used a set of 50,000 MRI brain scans from the NIH-supported Human Connectome Project to train the AUTOMAP system to reconstruct images in their study, successfully demonstrating improvements in reducing noise and reconstruction artifacts over existing methods.

AUTOMAP achieves almost instantaneous image reconstruction, according to senior author Matt Rosen, Ph.D., director of the Low-field MRI and Hyperpolarized Media Laboratory and co-director of the Center for Machine Learning at the MGH Martinos Center. The reason for the rapid processing speed — just tens of milliseconds — is that the neural network has no cycles or loops, rather is a feedforward system.

“Some types of scans currently require time-consuming computational processing to reconstruct the images,” Rosen said. “In those cases, immediate feedback is not available during initial imaging, and a repeat study may be required to better identify a suspected abnormality. AUTOMAP would provide instant image reconstruction to inform the decision-making process during scanning and could prevent the need for additional patient visits.”

“This technology could become a game changer, as mainstream approaches to improving the signal-to-noise ratio rely heavily on expensive MRI hardware or on prolonged scan times,” said Shumin Wang, Ph.D., director of the NIBIB program in magnetic resonance imaging. “It may also be advantageous for other significant MRI applications that have been plagued by low signal-to-noise ratio for decades, such as multi-nuclear spectroscopy.”

For more information: www.nature.com/nature

Reference

Zhu B., Liu J.Z., Cauley S.F., et al. "Image reconstruction by domain-transform manifold learning." Nature, March 21, 2018. https://doi.org/10.1038/nature25988

Related Content

Single-slice of the chest CT showing the abnormalities. Arrows indicate the location of the breast mass (red arrow), lymphadenopathy (blue arrow), and a lung nodule (yellow arrow). Arrows not present in experimental display. Image courtesy of Psychonomic Bulletin & Review

Single-slice of the chest CT showing the abnormalities. Arrows indicate the location of the breast mass (red arrow), lymphadenopathy (blue arrow), and a lung nodule (yellow arrow). Arrows not present in experimental display. Image courtesy of Psychonomic Bulletin & Review

Feature | Radiology Imaging | November 20, 2020
There’s a classic video demonstrating how our brains process information and allocate attention in which people bounc
Densitas Inc., a global provider of A.I. technologies for digital mammography and breast screening, announced its partnership with Mammography Educators to offer the first artificial intelligence powered telehealth technologist training platform to support business continuity in mammography facilities.
News | Mammography | November 20, 2020
November 20, 2020 — Densitas Inc., a global provider of A.I.
1H-MR spectra of 3 consecutive patients with COVID-19. Upper row: Axial FLAIR images at the corona radiata level show representative MRS voxels (black squares) from sampled periventricular regions. Lower row: Corresponding spectrum (black) and LCModel fit (red) from each patient acquired at TE = 30 ms (upper row) and TE = 288 ms (lower row). A, A patient with COVID-19-associated multifocal necrotizing leukoencephalopathy shows diffuse patchy WM lesions with markedly increased Cho and decreased NAA, as well

1H-MR spectra of 3 consecutive patients with COVID-19. Upper row: Axial FLAIR images at the corona radiata level show representative MRS voxels (black squares) from sampled periventricular regions. Lower row: Corresponding spectrum (black) and LCModel fit (red) from each patient acquired at TE = 30 ms (upper row) and TE = 288 ms (lower row). A, A patient with COVID-19-associated multifocal necrotizing leukoencephalopathy shows diffuse patchy WM lesions with markedly increased Cho and decreased NAA, as well as elevated Lac. B, A patient with COVID-19 after recent PEA cardiac arrest with subtle FLAIR hyperintense white matter changes also shows elevated Cho/Cr and decreased NAA/Cr ratios. However, these derangements are less severe than in the patient in A. There is no clear elevation of Lac. C, A patient with COVID-19 without encephalopathy or recent severe hypoxia has normal Cho/Cr, with mildly decreased NAA/Cr and no lactate elevation. Cho, Choline; NAA, N-Acetyl-Aspartate; mI, Myo-Inositol; Lac, Lactate; Glx, Glutamate + Glutamine. Image courtesy of AJNR

News | Coronavirus (COVID-19) | November 19, 2020
November 19, 2020 — One of the first spectroscopic imaging-based studies of neurological injury in...
In partnership with Imaging Technology News, we collected feedback from more than 100 diagnostic imaging professionals including healthcare executives, radiology and IT leaders, radiologists, technologists, clinicians, and industry vendors regarding their perspective of the new virtual format, whether they’re attending, and if so what they’re hoping to get out of it. #RSNA20
Feature | Radiology Business | November 17, 2020 | By Laurie Lafleur
SyntheticMR, an innovative software solutions company for magnetic resonance imaging (MRI) announced today the company filed a 510(k) submission for its musculoskeletal solutions package, SyMRI MSK.
News | Magnetic Resonance Imaging (MRI) | November 17, 2020
November 17, 2020 — SyntheticMR, an innovative software solutions company for...
Acquisition expands Medtronic's artificial intelligence and data capabilities, becoming the first company to offer an integrated spine solution including AI-driven surgical planning, personalized spinal implants and robotic assisted surgery

Getty Images

News | Artificial Intelligence | November 16, 2020
November 16, 2020 — Medtronic announced that it has completed its friendly tender offer for France-based...
Dia's LVivo artificial intelligence software can help automate many features of echocardiograms to speed workflow and aid novice users. The software is now integrated into the Konica Minolta Exa PACS. 

Dia's LVivo artificial intelligence software can help automate many features of echocardiograms to speed workflow and aid novice users. The software is now integrated into the Konica Minolta Exa PACS. 

News | Artificial Intelligence | November 12, 2020
November 12, 2020 – Konica Minolta Healthcare Americas Inc. and DiA Imaging Analysis Ltd.
Just as it has been difficult for us to predict the course of this pandemic, so too have healthcare organizations been challenged to predict and evolve their operations to optimize patient care — as well as revenue. In the initial wave, healthcare’s technology needs shifted rapidly.

Getty Images

Feature | Artificial Intelligence | November 12, 2020 | By Mark Dobbs
Just as it has been difficult for us to predict the course of this...