“Federated learning,” as it’s called, works by training an algorithm across multiple data servers, without exchanging any images, experts detailed in Scientific Reports. 

That’s according to the results of a recent survey conducted at one private Colorado imaging practice, highlighted in the American Journal of Roentgenology. 

New capabilities use interoperability to enable patients and physicians alike to instantly access medical imaging. 

It says more than 200 healthcare institutions across the globe have now expressed interest in participating, with “substantial” datasets already added to the repository. 

When a child shows up at the emergency department with a traumatic brain injury, the encounter can often result in a costly head CT scan that provides little benefit to the patient, experts wrote Tuesday in JACR. 

Participating institutions will contribute demographic information, signs and symptoms data, imaging exams, lab tests and patient outcomes.

That’s according to a new survey of various clinician types, published recently in the Annals of the American Thoracic Society. 

Others have previously designed such reporting and data systems, but those do not specifically address radiology department workflows, Brigham and Women's experts wrote in JACR.