The radiology department at New York University School of Medicine in New York City has joined forces with Facebook to launch fastMRI, a collaborative research project applying artificial intelligence to MRI scans.
It can take anywhere from 15 minutes to more than an hour to scan a patient in an MRI machine, which makes MRIs challenging for children, patients with claustrophobia and those with back pain, among other issues. With AI, NYU School of Medicine and Facebook think they can complete MRI scans up to 10 times faster.
The fastMRI project centers on changing the way an MRI machine creates an image. Today, an MRI puts together a rendition of a patient's internal organs by gathering a series of sequential views of a patient, which it translates into a cross-sectional image.
However, early research conducted at NYU School of Medicine suggests AI can generate high-quality images from limited data, meaning an MRI may be able to capture less data — and therefore scan more quickly — and still create an accurate image by applying this technology.
This method would rest on the researchers training an artificial neural network — which mimics the way a human brain learns and processes information — to recognize the underlying structure of the images it gathers. That way, it can fill in any omissions without sacrificing accuracy.
"A few missing or incorrectly modeled pixels could mean the difference between an all-clear scan and one in which radiologists find a torn ligament or a possible tumor," a Facebook blog post announcing the collaboration reads. "Conversely, capturing previously inaccessible information in an image can quite literally save lives."
To apply this early research to the medical imaging field, the research team will develop artificial neural networks using 10,000 clinical cases and 3 million MRI scans of knees, brains and livers collected at NYU School of Medicine. NYU School of Medicine said it has stripped all identifying patient information from the clinical data to ensure the work is HIPAA-compliant and approved under NYU Langone Health's institutional review board.
Facebook will participate in the research project through its research arm FAIR — or Facebook Artificial Intelligence Research — and supply the fastMRI team with its computer vision expertise and ability to train AI models at scale.
In April, CNBC reported that Facebook had asked several large U.S. hospitals to share anonymized patient data for a research project. At the time, Facebook said it was contemplating a project to match patient data with Facebook data to help hospitals determine which patients require extra care. However, the social network pumped the brakes amid the Cambridge Analytica scandal.
The fastMRI research project, by contrast, will not use any Facebook data, according to the blog post.
"This collaboration focuses on applying the strengths of machine learning to reconstruct the most high-value images in entirely new ways," Facebook's blog post reads. "Our aim is not simply enhanced data mining with AI, but rather the generation of fundamentally new capabilities for medical visualization to benefit human health."