Facebook Collaborates with NYU School of Medicine to Make MRI Scans Faster

The project will initially focus on changing how MRI machines operate. The researchers hope that it may be possible to capture less data and scan faster by using AI, while preserving or 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 the views omitted from the accelerated scan. Early work performed at the NYU School of Medicine has shown that artificial neural networks can generate high-quality images from far less data than was previously thought to be necessary.

The imaging data set used in the project has been collected exclusively by the NYU School of Medicine and consists of 10,000 clinical cases and comprises approximately three million magnetic resonance images of the knee, brain, and liver. Unlike other AI-related projects, which use medical images as a starting point and then attempt to derive anatomical or diagnostic information from them (in emulation of human observers), this collaboration will focus on applying the strengths of machine learning to reconstruct the most high-value images in entirely new ways. With the goal of radically changing the way medical images are acquired in the first place, the aim is not simply to enhance data mining with AI, but instead to generate fundamentally new capabilities for medical visualization in order to benefit human health.

Facebook plans to open-source the work to allow the wider research community to build on the developments. As the project progresses, Facebook will share the AI models, baselines, and evaluation metrics associated with this research, while the NYU School of Medicine will open-source the image data set. This is expected to help ensure the work’s reproducibility and accelerate the adoption of resulting methods in clinical practice.

The project will initially focus on MRI technology, although its long-term impact could extend to several other medical imaging applications. For instance, the improvements afforded by AI also have the potential to revolutionize CT scans. Advanced image reconstruction could enable ultra-low-dose CT scans suitable for vulnerable populations, such as pediatric patients. Such improvements can help transform the experience and effectiveness of medical imaging, as well as equalize access to an indispensable element of medical care.

“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,” stated Facebook.

Related Links:
Facebook
NYU School of Medicine’s Department of Radiology