Underreporting of findings on CT exams is relatively common, but a deep learning algorithm showed “superior diagnostic performance” in spotting two particular concerns. 

ACR said the updates pertain to neuroradiology and cover clinical scenarios including white matter lesion tracking in multiple sclerosis patients. 

The investigation was retrospective, but Duke scientists believe their algorithm could potentially aid radiologists in spotting near-misses in their work.

The Category 3 code proposal would cover AI analysis for the detection of vertebral fractures and could take effect as soon as Jan. 1, 2022. 

RadTranslate is a web app that creates AI-spoken, natural-sounding audio clips, and it's available free of charge for practices struggling to find interpreters.

DeepHealth, a subsidiary of the Los Angeles-based imaging giant, said its novel algorithm demonstrated higher performance than a group of trained breast radiologists. 

ABR said artificial intelligence will help determine which oral and computer-based exams to review more closely for infractions. 

The algorithm also demonstrated “excellent” performance in spotting lung cancers missed on chest x-rays, experts wrote in Radiology: Cardiothoracic Imaging.