Facebook, NYU School of Medicine Collaborate to Improve MRIs

Posted August 28, 2018

In an effort to improve healthcare technology, two very different organizations are teaming up—social media giant Facebook and the New York University (NYU) School of Medicine’s Department of Radiology. The two institutions have announced a collaborative research project to improve the speed of magnetic resonance imaging (MRI) scanning through the use of artificial intelligence (AI).

MRI scans currently take about 15–60 minutes to complete, compared to seconds-to- minutes for other imaging methods, such as computed tomography or X-ray. NYU and Facebook’s project, called fastMRI, allows for MRI scans that can be completed up to 10 times faster than current MRI methods. While the images produced don’t look like any image a human radiologist could read, the project aims to use AI methods to reconstruct a high-quality radiological image from the partial data that is obtained from a rapid scan.

“Using AI, it may be possible to capture less data and therefore scan faster, while preserving or even enhancing the rich information content of magnetic resonance images. The key is to train artificial neural networks to recognize the underlying structure of the images in order to fill in views omitted from the accelerated scan,” the researchers wrote in a press release. “In practice, reconstructing images from partial information poses an exceedingly hard problem.”

To study the method, a team of engineers, mathematicians, clinicians, and scientists at NYU’s Center for Advanced Imaging Innovation and Research, along with members of the Facebook Artificial Intelligence Research group will examine deidentified information from 10,000 clinical cases, including approximately three million MRIs of the knee, brain, and liver from the NYU School of Medicine.

By reducing scan time, the researchers said they hope to reduce scheduling backlogs in rural areas as well as improve comfort for claustrophobic patients, who find MRI scans difficult to endure. In addition, if MRI scans become faster and more accessible, then they may be able to replace imaging methods that expose patients to radiation, the researchers wrote.