Dear AuntMinnie Member,
Patient education has been considered an important potential application for ChatGPT and other large-language models in radiology. However, these types of chatbots may not be ready for that role yet if new research is any indication.
A study published this week concluded that ChatGPT and Google's Bard model were able to simplify responses to patient questions about lung cancer screening, but not to the optimal reading level. What's more, researchers also found that, in comparison with patient education content on the Society of Interventional Radiology's website, ChatGPT sometimes produced inaccurate or incomplete information.
AI and contrast-enhanced mammography
Stories on women's imaging topics were also popular this week, including an article describing how a deep-learning model can identify and describe suspicious lesions on contrast-enhanced mammography. Speaking of AI, another study reported that a deep-learning algorithm could differentiate between benign and malignant breast microcalcifications.
You can also find these stories and more, including coverage of research on optimizing surveillance imaging, in your Women's Imaging Community.
SalaryScan
Have you filled out your SalaryScan survey yet? If not, there's still time to contribute to our annual initiative to gather the latest compensation and benefits data for the radiology community. Your responses will be kept totally anonymous and will help us produce the most accurate results for you and your fellow AuntMinnie.com members.
