Artificial intelligence could improve rads' productivity but also decrease demand for their services, according to a new Association of American Medical Colleges report. 

Human providers typically spot as few as 30% of these conditions before birth, but a new machine learning model from UCSF bumped the rate up to 95%. 

The Plantation, Florida, imaging center operator declined to disclose its partner's name for "strategic reasons." 

Dutch researchers believe the program could lead to fewer unnecessary diagnostic interventions, lower radiologists' workloads and reduce the costs of cancer screening. 

Taking a radiologist-centric approach to AI development holds promise for catching systematically missed lung cancers, experts wrote recently. 

NYU scientists developed their computer program using more than 5,200 radiographs gathered from 3,000-plus critically ill coronavirus patients treated at the institution. 

The Santa Clara, California-based company has raised more than $49 million in funding and inked partnerships with providers including RadNet and the Mallinckrodt Institute of Radiology. 

Michigan Medicine scientists helped design a deep convolutional neural network that could pinpoint nearly 91% of occult fractures on scaphoid radiographs.

Medo-Thyroid processes video sweeps of the glands, with AI selecting optimal images, calculating measurements, and helping characterize nodules using TI-RADS. 

Leaders in the space shared their predictions for imaging AI during a panel hosted by the Stanford Institute for Human-Centered Artificial Intelligence. 

The finding is part of the the American College of Radiology’s Data Science Institute's inaugural annual survey, published in JACR. 

With the recent greenlight, the Tustin, California, company can now offer its Deep Learning Spectral CT tool for cardiovascular applications.