We have also developed a wonderful collaboration with the American College of Radiology (ACR) and the American Association of Physicists in Medicine (AAPM) on a large research image repository and associated tools, called the Medical Imaging and Data Resource Center (MIDRC). That initiative is funded by the National Institute of Biomedical Imaging and Bioengineering (NIBIB) and the Advanced Research Project Agency for Health (ARPA-H) and will facilitate AI imaging research for years to come.

HCB News: Has radiology's relationship with AI changed since the last RSNA meeting?
CL: Eight years ago, when everyone first realized the promise and power of AI, there was a certain amount of anxiety and fear about how it would affect the radiology workforce. As we gain experience with AI, we can see its power, but also its limits. And of course we have an ever-growing radiology work force shortage, so AI is viewed as part of the solution.
A few trends are clear now: Number one, radiology continues to lead the way in AI, with more than three-quarters of FDA cleared AI products targeting radiology. Number two, AI algorithms are playing a larger role in diagnostics, performing detection and measurement tasks that are tedious or ill-suited to the human visual perception system. And number three, many of the earliest radiology AI products focused on detection, which don’t provide much of a productivity advantage and can even slow us down—not very appealing to radiologists. So there is a renewed focus on opportunistic screening and on work flow applications like reporting tools and medical record summarization, which have a clearer return on investment.

I am also excited about what I call the coming “ChatGPT moment” in medicine. ChatGPT was trained on massive amounts of data. But we haven’t yet seen medical applications trained on similar amounts of data. When we do, I think we will see some surprising capabilities, just as ChatGPT surprised us. So I am optimistic about the next wave of radiology AI applications.

HCB News: Are there events or presentations you are most looking forward to at this year's meeting?
CL: I will be talking about all these AI issues during my presidential address at the Opening Session on Sunday afternoon. The theme of this meeting is Building Intelligent Connections. So I will highlight how we can connect with technology and with each other to spark innovation and optimize patient care. I will use a patient story told twice, once pre and once post AI, to show how AI and other high-tech tools can improve the experiences of patients and radiologists.

More Exclusive Q&A

More Voices