What 2 former FDA officemates wish you knew about radiological health regulation

They spoke with RAPS to explain why radiological health often gets overlooked, why so many people need to understand it, what they wish they knew about radiological health back when they started at the FDA and more. This conversation has been lightly edited for clarity.
 

RAPS: Jay, you recently wrote that most people don't realize how radiological health impacts their lives, literally from cradle to grave. Imaging is used during pregnancy. Perhaps a broken bone needs to be X-rayed later in life. Cancer is detected by way of medical imaging exams. MRIs are even used after death, during autopsies. What's even more curious is that you wrote that most people, even people who should know, like regulatory professionals and doctors, don't necessarily understand how these radiological health devices even come to market. Why is there such a gap in that knowledge and how does this new book help close that gap?
 

Jay: I think medical imaging and regulatory are both two things that, when they're working well, nobody thinks about them because they don't have to. Ultrasounds, for example. When you're pregnant and you have your 20-week ultrasound, you're thinking, “Is my baby a boy or are they a girl?” You're not thinking, “How did the ultrasound transducer come to market?” At least most people aren't thinking that. But, like I said, when these things work, well, people don't think about them. And really, that's what we hope this book does, is to get people thinking. We trust, from cradle to grave, so much of our medical care and lives to these devices working properly. And it's really important that people understand, when we trust these devices, what has happened to make sure that they're trustworthy? And I think that's what our book tries to explain.
 

RAPS: So, From X-Rays to AI is the first book of its kind. What are some of the other topics that this book digs into that others don't? And why does radiological health seem to get the short shrift when it comes to this type of education?
 

Jay: Interesting question. I teach classes on regulatory. And when you do regulatory education, you tend to aim for a very broad audience. You teach about devices or drugs and not really about a specific device. And there are some things, broadly, that apply to many devices, like biocompatibility or sterility or software. But in radiological health, you have a lot of unique issues like the dose of radiation that's delivered to patients or the use of radiopharmaceutical drugs, which are drugs that are radioactive sources themselves. These issues are very important to medical care, but they're specific to radiological health. And a lot of people who are preparing broad education on devices may not have background in this area. That's really, I think, part of the reason that it doesn't make it into general device courses.
 

RAPS: Who would benefit from reading this book? Obviously, anyone involved in regulatory work around these devices or those who want to better understand the regulatory landscape would find it to be super useful. But who else would find this book to be essential reading?
 

Jay: I think if I were a healthcare provider and I were using these devices, if I were radiologist or radiation oncologist or even a CT tech or a medical physicist, I'd kind of want to know: this device that I use on patients, how did it come to market? What sort of testing was done on it? What was the process by which this device came to my hospital? If I were interested in how many AI-based medical imaging devices are coming to market, I'd want to read it too. If I were starting a startup or something like that, I think this is helpful. It gives an overview of the field. So, these are just some of the things that come to mind. And of course, if I were a patient and I had questions about, "What was done to make sure that this device is safe and effective?” I'd be interested, too.
 

RAPS: So, Jay, what are some key takeaways from this book and what are some of the things that people would be surprised to learn?
 

Jay: Radiological devices are in Class II, mostly. Not all of them, but most of them. And people therefore think, okay, Class II, you just have to prove substantial equivalence. So, it must be straightforward. But actually, I think what's surprising is to see how non-straightforward it is. And why is that? Because we're talking about radiological images. So much depends on the quality of the image. And the quality of an image is hard to measure. Just like if I ask you, is a photo good or bad? That's hard to answer because it means different things in different contexts to different people.
 

What would surprise readers? I think I was surprised to learn that the major postmarket episode involving a nuclear medicine product had nothing to do with radiation and instead involved a gamma camera falling on a patient. Additionally, we think of AI as this new thing, but it has a very long history in mammography.
 

RAPS: Both Yuan and Jay are ex-FDA officials. How did your time at the agency inform the work that you did on this book?
 

Yuan: A fun fact about Jay and I is that we were actually officemates at the FDA. A lot of our colleagues were on the same floor, so we learned so much from each other in terms of expertise and how to handle reviews and questions. This book really distills our learnings both from the agency and industry, and that's what we really were excited about.
 

RAPS: I found it interesting that both of you worked in very close proximity at the FDA.
 

Jay: Yeah, within a couple of feet of each other in a small FDA office. Our FDA time, it's really influenced me. It taught me regulatory, but then when I moved outside of the FDA it taught me that the FDA perspective is important, but it's only one piece of a broader landscape. That’s why we recruited authors not only from FDA, but from industry and from consulting, so that the whole field is really represented.
 

RAPS: What are some of the more important things that the FDA wants to see from manufacturers of radiological health devices when it comes to complying with agency rules? And is there a particular aspect of this topic that perhaps the FDA regularly sees companies falling down?
 

Jay: I think, in general, the fundamental philosophy of FDA is that the data that you provide should support whatever marketing claims that you're trying to make about your device. Where a lot of companies fall down is just not having that support. Part of it is because image quality is so important in the field of radiological health. And image quality is a subjective and slippery thing to measure. The same image can have different qualities if you're doing different things with it, because the quality is really, how well can a doctor do what they need to do using the image? And different doctors can be doing different things. So yeah, so it's really image quality and how to measure image quality.
 

RAPS: What's interesting is that this book, From X-Rays to AI, could be used by the FDA as they work with new agency recruits to bolster the understanding of these types of devices. Why might FDA find this to be such a useful tool for new hires?
 

Jay: So, when I first joined FDA, my boss told me it takes people about two years to feel confident as a reviewer. And when my two years was up, I thought about, why did it take that long? Part of the reason for that huge learning curve was that there was very little written about the regulation of radiological health and certainly there was nothing that collected all of it in one place and nothing that included perspectives from industry as well. 

This book is the resource I wish I had when I started. And even as I grew and went on, once I became specialized on a particular device, it was very easy to get tunnel vision and not know what was going on in the rest of the space. This book is set up so that a reader can dip into a chapter and learn about an individual area of radiological health.
 

Yuan: The COVID experience really informs my perspective. Compared to the days where Jay and I were officemates, folks with hybrid or remote work situations have a lot less opportunities for face-to-face discussions and learnings. And that's why the intention of this book is to allow us to be able to distill all learning for future recruits and future reviewers at the FDA.
 

RAPS: This is the first time that either of you have authored a book for RAPS. So, why this book? What is it about radiological health that got you both off the bench, as they say, and into the game?
 

Yuan: Jay and I both have PhDs. We have published book chapters and journal proceedings. But I think we're at a point where we can distill our FDA and industry experience to make this contribution, and this is what excites me about this book.
 

Jay: I feel similar to Yuan. Also, when I was studying for the RAC exam, I used the RAPS Fundamentals book and I really thought, “Oh, it'd be fun to write one of these.” And my expertise was radiological health, so that was the natural fit. I've spent most of my career working with radiation in some form or another. I trained as a physicist, and so, you know, it wasn't really a big jump that this would be the subject.
 

RAPS: Before we close, one last question: and that is tell us what you think someone will get out of purchasing your book, From X-Rays to AI? Why should they buy it?
 

Yuan: From my perspective, this is a one-stop shop. We have seasoned authors from industry, including manufacturers, law firms, and also current and former FDA experts contributing to this book.
 

Jay: I think everybody right now wants to know about artificial intelligence and machine learning in medicine. And radiology is the big place where it's really being deployed. We've tried to touch on it in every single chapter of this book and give you some case studies of products that have made it to market. And also, what are some ways that AI is being applied, apart from just the obvious one, which is computer assisted diagnosis?

From X-Rays to AI: Navigating US Regulations in Radiological Health is available now. Get your copy today.