Health IT, Hospitals

Facebook, NYU School of Medicine team up for AI MRI research

Called fastMRI, the research collaboration will examine how artificial intelligence can make MRI scans up to 10 times faster.

Six panels of an MRI scan.

Social media giant Facebook and NYU School of Medicine’s Department of Radiology have collaborated to examine how artificial intelligence can boost the speed of MRI scans.

The project, dubbed fastMRI, will focus first on altering how MRI machines operate.

Usually, scanners gather raw numerical data and turn it into cross-sectional images of internal body structures. When more data has to be captured, the scan takes longer. Overall, the scan can take anywhere from 15 minutes to more than an hour. But by deploying AI, the organizations hope to gather a smaller amount of data and therefore scan faster.

NYU School of Medicine collected the data that will be used in fastMRI. It includes 10,000 clinical cases and comprises about 3 million images of the liver, knee and brain. The work is HIPAA-compliant, and patient names and other PHI have been removed from the data. No Facebook information will be used.

Though the initial goal is to make MRI scans up to 10 times faster, Facebook and the medical school believe their work could also impact other medical imaging applications like CT scans.

“We believe the fastMRI project will demonstrate how domain-specific experts from different fields and industries can work together to produce the kind of open research that will make a far-reaching and lasting positive impact in the world,” reads a Facebook blog post on the announcement.

NYU School of Medicine’s Department of Radiology has long been involved in improving medical research, particularly through its Center for Advanced Imaging Innovation and Research, which focuses on rapid image acquisition and advanced image reconstruction, among other topics. The center has been exploring how to use AI for faster MRI scans since 2016.

Meanwhile, the Facebook Artificial Intelligence Research group, which tackles advancing AI, was looking for ways to make a real-world impact. The opportunity with NYU popped up and the collaboration was born.

The Menlo Park, California-based company has been inching its way into the healthcare landscape in other ways.

Earlier this spring, CNBC reported Facebook was in talks with U.S. hospitals to share anonymized patient data for a research project, but the plan was put on hold. The fact that the project was halted was hardly surprising, given that Facebook was in hot water following the Cambridge Analytica scandal.

At its F8 developer conference this year, the social media company announced a new capability that allows users in India, Bangladesh and Pakistan to find nearby opportunities to donate blood.

Photo: maciu17, Getty Images

Shares1
Shares1