A recent study led by Jonathan Huang at Northwestern University's Feinberg School of Medicine demonstrated that radiologists utilizing generative AI improved their radiograph reporting efficiency by over 15% without compromising diagnostic accuracy. The research included 23,960 radiographs and found that AI-assisted reporting resulted in a mean documentation time of 159.8 seconds compared to 189.2 seconds for traditional methods. Furthermore, the AI model showed a specificity of 99.9% in identifying clinically significant pneumothorax and flagged priorities much faster than human notification.

Source: JAMA Network Open