Mass General researchers develop AI technique to speed image reconstruction in radiology

  • Small
  • Medium
  • Large

A team of researchers from Massachusetts General Hospital in Boston developed an artificial intelligence technique to improve the quality of medical images, the hospital announced March 21.

To generate a high-quality medical images, clinicians need to acquire sufficient data. This task often comes at a cost to the patient, including an increased radiation dose for CT or PET scans and elongated scan times for MRIs. However, the AI-powered technique may enable clinicians to create high-quality images without collecting additional data.

"An essential part of the clinical imaging pipeline is image reconstruction, which transforms the raw data coming off the scanner into images for radiologists to evaluate," said Bo Zhu, PhD, a research fellow in the hospital's Athinoula A. Martinos Center for Biomedical Imaging and first author on the research team's paper, which was published in Nature March 21.

Using machine learning, the imaging system — dubbed automated transform by manifold approximation, or AUTOMAP — identifies appropriate image reconstruction algorithms to produce accurate images. The system, which was trained using real-world medical images, reconstructs an image in tens of milliseconds, according to the hospital's March 21 statement.

The accelerated processing speed may prove useful for clinicians looking to make real-time decisions about imaging and care protocols while the patient is being scanned.

"AUTOMAP would provide instant image reconstruction to inform the decision-making process during scanning and could prevent the need for additional visits," said senior author Matt Rosen, PhD, director of the Low-field MRI and Hyperpolarized Media Laboratory and co-director of the Center for Machine Learning at the MGH Martinos Center.

Copyright © 2021 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.


Featured Whitepapers

Featured Webinars