Hugh Harvey is a consultant radiologist in the UK and has acted as an advisor to AI startups, including Kheiron Medical.
Amy Patel is a breast radiologist at Beth Israel Deaconess Medical Center in Boston, Massachusetts. She is also on the steering committee for RADxx, an initiative to recruit women into imaging informatics and to foster mentorship and networking opportunities for women radiology leaders.
FHI: How is technology changing the way a radiologist must train?
Hugh Harvey: I see the role evolving and changing, but technology can’t replace the meaningful work that a radiologist does. Some things won’t change – radiologists still need to learn the human anatomy and obtain the exact same medical training as any other doctor.
What will change, though, is the added need to be technologically proficient and have an understanding of the statistical underpinning of the technology. So if you’re interacting with AI, you’ll need to understand the typical analyses and have an understanding of probability, for example. So there will be a mathematical and technological element added to the radiologist’s curriculum.
Amy Patel: In the US, we are seeing a lot of interest from trainees desiring additional training. The American College of Radiology hosts an online Resident and Fellow Section Journal Club, which has been the most attended of all the virtual journal clubs it’s ever hosted. A recent article published in Academic Radiology by radiologists at Massachusetts General Hospital points out that radiologists are taking on the responsibility of using AI tools on top of image accusation and interpretation, so we as educators are going to have to decide which AI tools should be introduced to trainees and at what time. I wholeheartedly agree with this.
FHI: How much of the drive for new technology is coming from below – from the younger generation of trainees?
AP: I’m just a first year attending out of training and my generation is definitely more adaptable to new technologies. I think that there is a lot of potential to really change the landscape in radiology and the good news is that the next generation of radiologists will be ready to embrace it and lead the charge.
HH: I think that’s true. The American College of Radiology is doing phenomenal things with its journal club. I don’t know of any other country that’s doing so well in terms of being led by the trainees.
FHI: How important is it for a radiologist to be a part of multidisciplinary activities in the hospital?
AP: It’s more relevant now than ever. In the US, there’s an initiative called Imaging 3.0 that places an emphasis on the shift from volume- to value-based care. And as part of this value-based system, multidisciplinary activities are crucial. It’s an absolute disservice to the patient if we’re not all on the same page providing coordinated and cohesive care. For example, I’m a breast radiologist and it’s imperative that I sit down with the oncologist, the surgeon, the pathologist and nurse navigator. And the multidisciplinary approach is relevant now more than ever as we see this new push in healthcare for personalized medicine and interactions. Oftentimes, the rest of the care team can help put the pieces of the puzzle together, in terms of patient desires regarding management and other factors, but having this comprehensive insight can be instrumental in how we approach the tailored care of each and every patient. Radiologists are realizing that we have to come out of our dark rooms and interact with the rest of the team. At the end of the day it’s the best thing for the patient.
HH: It’s a little different in the UK. We do have multidisciplinary meetings, but they’re mainly for the consultant level. But I think that technology will actually enable radiologists to spend more time engaged in this kind of activity, so I’m all for it.
FHI: Is there a risk that technological innovations are commoditising imaging as a technical service and could undervalue the expertise of radiologists?
HH: Yes, there’s definitely a sense that radiology can be seen by other medical specialists as a kind of black box where they put in a request for an order for a scan and they simply get a result out.
The 1970s have become known as this ‘golden era’ of radiology, where radiologists were the hub of the hospital. Other doctors would do their rounds, end up in the radiology department and have long discussions about different cases. Now everything is digitized and simplified, which is why the radiologists have sort of been forced into the dark where no one goes to interrupt them anymore.
I would really like to see radiologists step into the light and become more patient-facing, letting AI take the burden of some of the routine stuff. Otherwise we could end up as outsourcing departments in the middle of nowhere, which I don’t think anyone wants to see happen.
AP: It’s interesting that Hugh mentioned the ‘golden era’ of radiology because that is brought up a fair amount in the US. And now there are misconceptions that radiologists are not playing as integral a role in patient care.
We need to make it clear that we are available for referring physicians to discuss the diagnosis and treatments of patients, including complex cases that demand a team effort.
In breast imaging, discussing a result with the patient is central to what I do, particularly with diagnostic examinations, and it puts patients’ minds at ease, especially if they have had a previous cancer diagnosis. So as much as AI and all this technology has the potential to create pretty significant changes in radiology, these intangibles that I speak of and the human touch cannot be replaced. So radiologists have to reiterate that we can add value where technology cannot.
HH: However, a knee surgeon is actually already very good at looking at imaging of the knee. If we have an AI system that can potentially provide him or her with all the information to make a decision on what to operate on, then the radiologist is not really required. This could be the danger with AI – more work could move away from us. It’ll start to happen slowly in very specific niche areas.
AP: Yes, there is at least a risk of us becoming less relevant, but again it comes down to us stepping out of our dark rooms so that we aren’t just known as a source that can be dictated and driven by AI and machine learning.
FHI: So what can radiologists do to ensure their expertise remains valuable in modern healthcare?
HH: We have to provide a robust and consistent service that is going to help physicians find the answers that they want. Radiology is also becoming much more consumer-facing. Patients appreciate it when a radiologist explains their imaging to them directly – they feel more comfortable under the radiologist’s care. I think patient-facing initiatives are going to be much more important.
AP: I really think that Hugh has hit the nail on the head in so many respects. In the US we are transitioning from volume to value, which is going to ultimately dictate how we are reimbursed and the future payment models in this country. We must continue to be present and stress that this paradigm shift will not be achieved if we as radiologists are not present and more patient facing, providing tangibles and expertise that technology cannot. With the advent of artificial intelligence, this will only improve what we do as radiologists, but ultimately I feel we will always remain relevant. Providing these other intangibles – that’s how we can continue to be incredibly valued within the care of the patient.
HH: In the UK, the National Health Service (NHS) needs to embrace AI across the board and help build it and actually validate it. For example, one of the companies I work at can detect cancer and other cells before humans can see them, so radiologists need to embrace these technologies and buy these services as part of their package of care.
I think that radiologists can certainly incorporate AI in their practice. At the moment, many of us are overwhelmed with information and we need AI to help us find the right information at the right time.