Facebook (Menlo Park, CA, USA) and NYU Langone Medical Center (New York, NY, USA) have created an artificial intelligence (AI) system that can speed up magnetic resonance imaging (MRI) examinations.

The new system, called fastMRI, is based on a deep learning (DL) image reconstruction model optimized via dedicated multi-sequence training, in which a single reconstruction model is trained with data from multiple sequences with different contrast and orientations. Following training, data from 108 patients were retrospectively undersampled in a manner that would correspond with a net 3.49-fold acceleration of fully-sampled data acquisition and 1.88-fold acceleration. The ability of six readers to detect internal derangement of the knee was then compared for both clinical and DL-accelerated images.

The results demonstrated a high degree of interchangeability between the standard and DL-accelerated images; in particular, results showed that interchanging the sequences would result in discordant clinical opinions no more than 4% of the time for any feature evaluated. Moreover, the accelerated sequence was judged by all six readers to have better quality than the clinical sequence. The new technology also allowed creation of MRI films in 75% less time (about 15 minutes). The study was published on August 8, 2020, in the American Journal of Roentgenology.

“This study is an important step toward clinical acceptance and utilization of AI-accelerated MRI scans, because it demonstrates for the first time that AI-generated images are indistinguishable in appearance from standard clinical MR exams and are interchangeable in regards to diagnostic accuracy,” said lead author Professor Michael Recht, MD, of NYU Langone. “The results represent the culmination of nearly two years of open research from Facebook AI and NYU Langone’s fastMRI initiative, a collaborative effort to improve medical imaging technology and advance research using AI to generate images from limited data.”

Deep learning is part of a broader family of AI machine learning methods based on learning data representations, as opposed to task specific algorithms. It involves CNN algorithms that use a cascade of many layers of nonlinear processing units for feature extraction, conversion, and transformation, with each successive layer using the output from the previous layer as input to the new one in order to form a hierarchical representation.

http://www.gzjiayumed.com/en/index.asp