As these models continue to advance, it is important to understand how their performance holds up over time.

“When we rely too much on whatever the computer tells us, that’s a problem."

By expediting communication between radiologists and referring providers when suspicious findings are identified, the system has great potential to improve clinical outcomes in real-world scenarios, authors of a new study in JACR suggest.

Such technology can be leveraged to improve risk assessments without accruing additional expenses for patients or clinicians, potentially filling in gaps where DXA exams fall short.

Many of these files are still readily available online, and numerous contain PHI that is “easily accessible,” a new analysis reveals.

Three factors are associated with true high-risk findings—imaging protocols for aortic valve implantation planning, imaging area, and cardiology department orders.

The large language model can also outperform other human providers, radiologists included, new study shows.

With proper prompts and appropriate context, chatbots can play a valuable role in imaging referrals.