NEW YORK, Jan. 18, 2023 /PRNewswire/ — Artificial intelligence (AI) can reconstruct sampled rapid MRI scans into high-quality images with comparable diagnostic value to those generated by conventional MRI scans, according to a report. For a new study by New York Grossman School Medicine and Meta AI Research.
The study says that using artificial intelligence to reconstruct MRI scans that are four times faster than standard scans promises to expand access to MRI scans to more patients and reduce wait times for appointments.
The study, published January 17 in the rayspart of fastMRI initiative Founded by NYU Langone Health and Meta AI Research (formerly Facebook) in 2018. This initiative, which aims to use AI to make MRI scans faster, resulted in an AI model jointly developed by Meta AI researchers and imaging scientists and specialists at NYU Langone. It also produced the largest-ever publicly available set of raw MRI data, which has been used by scientists and engineers around the world.
In an earlier “proof-of-principle” study, Langone’s team at NYU simulated fast scans by removing about three-quarters of the raw data obtained with slow conventional MRI scans. The FastMRI AI model then generated images identical to those generated from the slow-motion scans. In this new study, the researchers performed quick scans using only a quarter of the total data and used an AI model to “fill in” missing information, similar to how the brain constructs images by filling in missing visual information from the local context. and previous experiences. In both studies, rapid MRI scans were found to be as accurate as conventional scans, with better quality.
“Our new study translates the results from the previous laboratory study and applies them to actual patients,” he says. Michael P. Recht, MDLouis Marx Professor of Radiology and Chair x_ray place at New York University Grossman School of Medicine. “FastMRI has the potential to dramatically change the way MRI scans work and increase the accessibility of MRI to more patients.”
In the new study, a total of 170 participants received an MRI of the knee using a conventional MRI protocol followed by an accelerated AI protocol between January 2020 and February 2021. Each scan was reviewed by six musculoskeletal radiologists, who looked for signs of meniscus or lace. Bone marrow rupture or cartilage abnormalities. The radiologists who evaluated the scans of the AI-reconstructed images were not told, and to reduce the possibility of retrieval bias, evaluations of standard images and AI-accelerated images were spaced at least four weeks apart.
The radiologists considered the AI-reconstructed images to be diagnostically equivalent to conventional images for detecting tears or abnormalities, and found the overall image quality of the quick scans to be significantly better than conventional images.
“This research represents an exciting step toward translating AI-accelerated imaging into clinical practice,” says Patricia Johnson, PhD, assistant professor in the department of radiology at New York University Grossman School of Medicine. “It really paves the way for more innovation and progress in the future.”
Expanding access to MRI
The researchers note that FastMRI requires no special equipment. Technicians can program standard MRI machines to collect less data than is normally required, and the fastMRI initiative has made its data, models, and code available as an open source project to other researchers, as well as manufacturers of commercial MRI systems.
With fastMRI, an MRI scan that may take up to 30 minutes can be done in as little as 5 minutes, making an MRI scan’s time comparable to an X-ray or CT scan. However, unlike these imaging techniques, MRI provides incredibly rich information, from enabling views of soft tissues from different perspectives to highlighting small cartilage abnormalities to locating tumors in the abdomen.
“The price we have traditionally paid for all this rich information is time,” he says. Daniel K. Sodikson, MD, PhDHead of Innovation in the Department of Radiology and Director Center for Advanced Imaging Innovation and Research. “If we can increase the speed of MRI to approach the speed of CT scans, we can make all of this important information available to everyone.”
This study was funded by National Institutes of Health grants R01EB024532 and P41EB017183.
View original multimedia download content:
Source: NYU Langone Health