The results can significantly expand access to MRI, improve the patient's experience, and potentially enable new use cases for MRI.
The new research validated the effectiveness of a novel AI (artificial intelligence) algorithm that generates accurate and faster MRI scans than the traditional MRI. Magnetic resonance imaging (MRI) is an imaging technique used in radiology to form pictures of the anatomy and the physiological processes of the body. The research was published in the American Journal of Roentgenology. The new algorithm generates accurate MRI results using around four times fewer data than usual. This innovation will help speed up MRI results without the need to upgrade already existing imaging hardware.
Moreover, rapid MRI scans produced with AI are as effective as a traditional MRI. The results can significantly expand access to MRI, improve the patient's experience, and potentially enable new use cases for MRI. For the study, a team of radiologists from Facebook's Artificial Intelligence Research (FAIR) and NYU Grossman School of Medicine has developed a neural network that can produce accurate and effective MRI scans than the traditional imaging technique. The team used the world’s largest open-source dataset of DE-identified knee MRIs to train the newly developed system.
It was developed by NYU as part of the FastMRI initiative and launched two years ago with Facebook. The AI ​​model generates a FastMRI scan that can match the scans produced by the standard MRI techniques. Now, individuals need to spend less time and can be imaged much faster as it requires four times fewer data than the traditional imaging technique. The team used both standard MRI imaging and the new AI model to create two separate datasets and determined the efficacy of the FastMRI. The team then evaluated both datasets and found that the newly developed AI model as effective as a traditional imaging technique.