From a central role in the COVID-19 pandemic to artificial intelligence advancements, CT further secured its role as a valuable arrow in the radiologist’s quiver.
Without a doubt, CT took center stage this year as radiology’s initial – and greatest – tool in the fight against the COVID-19 pandemic. Use of the chest CT scan was controversial internationally, but as the outbreak has raged on, the industry has discovered the value these images bring.
Still, the viral outbreak was not the only health need this year, and to meet the challenges presented by other conditions, providers and vendors expanded the way CT is being used, as well as the technologies that impact efficiency, dose, and workflow. Diagnostic Imaging spoke with Elliot Fishman, M.D., professor of radiology at Johns Hopkins Medicine, about the main developments in CT over the past year. Ultimately, he said, every advancement is making CT all the more valuable.
Diagnostic Imaging: This year has been a huge year for CT with the pandemic as the constant underlying issue. For the modality overall, though, what has been most significant in 2020?
Fishman: Obviously, in the beginning of the COVID-19 pandemic, it looked like everybody was going to be screening CTs for COVID-19 patients, and then it moved to just a select group of patients for whom it is probably necessary. We’ve been able to pick up many COVID-19 cases that were unexpected. This was certainly more of an issue before testing became more widely available.
But, I think CT has been good for looking at complications, whether it’s pulmonary, vascular, or gastrointestinal bleeding. CT has been very strong because COVID-19 patients are very complicated. Obviously, 80 percent of patients are fine, but that 10 percent or 20 percent that have issues actually have significant issues, and CT has been very strong at looking for complications and in managing the patients.
CT has also been helpful with a lot of sick patients who don’t have COVID-19. It could be surgical complications, post-operative things, or it could be appendicitis, diverticulitis, or myocardial infarction. The world keeps going on around COVID-19. There are a lot of diseases out there.
The second thing is artificial intelligence. A lot of great work has been done with artificial intelligence and recognizing COVID-19 pneumonia. When you look at artificial intelligence, almost all the work is on CT. Whether it’s intracranial bleeds, stroke, coronaries of the liver, the pancreas, or the kidneys, everything revolves around CT. It has really made CT even stronger because it’s the source of information, and you realize with artificial intelligence and things like radiomics that there’s so much information that you may not really appreciate, as the radiologist, that the computer appreciates.
So, CT, if you can imagine, is actually more valuable than ever, and it’s going to be even more valuable because there’s a lot of work going on with radiomics. CT can predict outcomes, not just simple staging. It can tell us who does well and who doesn’t do well. It can predict what chemotherapy should be used and it can help decide on surgical candidates.
AI really is the buzzword. Everything is artificial intelligence when you look at CT – whether it’s the acquisition of data, processing data, or streamlining when patients get read first. Reading with artificial intelligence is becoming almost ubiquitous. You’re not going to say you’re using artificial intelligence, it’s just going to be there. Artificial intelligence really is the biggest thing in CT. People aren’t talking about detectors. You don’t hear that discussion anymore. Now, it’s really what information you’re getting out of a CT and how you can use artificial intelligence to do more. I think that’s really exciting.
Diagnostic Imaging: Delving deeper into artificial intelligence, are there certain areas where the industry has seen the most development or places where the tools have provided new information that we have not seen before?
Fishman: The simplest one is using artificial intelligence for triage. There have been several FDA approvals, such as from Aidoc or Zebra, where the tools can simply look at the brain to determine if a bleed is present. If it looks like a bleed, that moves the patient to the front of the list. It shows you where the bleed is – so it’s been as accurate and more accurate than radiologists.
But, it’s more about just saving time. So, you’re going to read all these cases eventually, but you can save 40 minutes by moving it to the front of the list. Things like that become really critical – it’s a very practical thing.
People are using artificial intelligence for detecting pneumonia or lung nodules, for quantifying the degree of cancer. It is also being used for the early detection of pancreatic cancer, and people are using radiomics to determine grades of pancreatic tumors. It’s really going from an abstract process to a practical process that’s being used in clinical practice.
Diagnostic Imaging: What has been the impact of clinical practice thus far?
Fishman: I saw an article accepted to Radiology just the other day where they were using artificial intelligence to triage patients in the emergency room setting on CT once it’s done for managing patients. Some hospitals, some centers are already using it. It’s one thing to say that artificial intelligence is theoretical, and for most people, it is theoretical. But, there are a number of institutions that are already putting it in practice. So, I think it is very much on the way.
Diagnostic Imaging: In addition to artificial intelligence, what advances did the industry make with CT during 2020?
Fishman: All manufacturers have been showing more of a simple and easier interface for scanning where the computers are making more of the decisions rather than the technologists. Even the simplest things from centering patients in the gantry to deciding on the scan protocols. You get a quality study, but you minimize the dose. What we’re seeing is the use of artificial intelligence to make many of the decisions that were left up to technologists before. It’s going to be able to enhance the quality of the studies, again, while minimizing the dose.
Related Content: AI and Clinical Data Can Help CT Diagnose COVID-19
You’re also seeing increased work in post-processing. 3D imaging and cinematic rendering is a good example. That’s moving along. In addition, there’s the use of things like the HoloLens from Microsoft that is a way of improving visualization. This is a major push where we’re using technology from other fields to move into CT and help push discovery along there. For example, with something like voice to control and try to optimize interaction with the patients. You’re seeing a lot more of artificial intelligence in things like scheduling. I think the days of calling up and scheduling on the phone, for the most part, are ending because you can do things online. People like the idea of using OpenTable for dinner reservations, and they want some equivalent like that in radiology. They want the experience of medicine to be transparent. They want it to be easy-to-use. People have always complained that medicine is difficult, and they are tired of hearing that medicine is complicated.
Diagnostic Imaging: What can the industry anticipate going forward into 2021—what is on the horizon for CT?
Fishman: Again, artificial intelligence is going to go from theoretical to practical. That will be one of the biggest things whether it’s choosing contrast material or anything else. Vendors are also re-designing scanners to really minimize the technical input by technologists. Much more is becoming automated with decision-making. Before, so much more was left up to the technologist. That’s going to help decrease dose and increase throughput. And, as always, there’s a shortage of technologists, so that may help in that regard, as well. The challenge there, though, will be the adoption of new technology.
I would say one of the biggest threats to radiology is the trend of these gigantic groups. Now, you have these big groups of 2,500 people – they’re 2,500 employees, and they’re just following the rules that come down from above. How are you going to do medicine when you’re just an employee and not the owner and don’t have a voice in the decisions? What impact is that going to have when the interest in these big companies is not necessarily buying the newest equipment, the newest scanners, but maybe it’s making the scanners last a few years longer? How is that going to impact people in terms of how well they work? Is it going to foster a mindset where people just think of themselves as employees who read so many scans and get paid X amount of money? This could change the way people behave – it’s just human nature, and it could be a real challenge.
For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-Newsletter here.