Get the most out of your radiologist

Advice for hospitalists from a radiologist on how to choose the right scan for every patient.


Radiologists really do want to help hospitalists, radiologist Timothy P. Kasprzak, MD, MBA, assured attendees at Hospital Medicine 2017.

“We don't go into radiology saying, ‘Boy, I really want to say I can't exclude anything’ and giving you a bunch of differentials and basically saying everything is nonspecific,” said Dr. Kasprzak, who is director of abdominopelvic and oncologic imaging at Case Western MetroHealth in Cleveland.

But in order to provide specific reports, there are some things that radiologists need from the physicians who order scans. Dr. Kasprzak offered an overview of these common issues during his talk, “CT to PET Scans: What Every Hospitalist Needs to Know,” as well as a number of satirical headlines from GomerBlog, such as “Master radiologist able to hedge on every possible medical condition” and “Bored radiologist clinically correlates everything himself.”

Timothy P Kasprzak MD MBAslashPhoto courtesy of Dr Kasprzak
Timothy P. Kasprzak, MD, MBA/Photo courtesy of Dr. Kasprzak.

The path to not receiving “correlate clinically” as a result begins with ordering the right scan, according to Dr. Kasprzak. “There are some studies that are better for certain things than others. Computed tomography may be really good to look at things like free air, but it's really bad at looking at hepatobiliary, it's really bad at looking at prostate,” he said.

The best tool for choosing an imaging modality is the American College of Radiology (ACR) appropriateness criteria, according to Dr. Kasprzak. The criteria, developed by a committee of radiology experts, give clinical scenarios and evidence on which scan will be most effective, as well as the radiation dose associated with the test. “You're getting all the evidence of why the committee came up with this recommendation,” he said.

If that sounds like too much work, there's another option. “If you're not going to dig into the appropriateness criteria, ask a radiologist,” said Dr. Kasprzak. “If you have a complex case … you're not quite sure what studies would be the best, just ask a radiologist.”

Routinely sharing case details with the radiologist can improve the clarity of imaging results. “I've actually seen ‘rule out pain’ as a history,” he said. “It's OK if you're fishing, that's all right, but just kind of own it and say that: ‘I don't know what's going on. Vague nonspecific abdominal pain.’”

Selection of an imaging test also often involves decisions about contrast. Contrast adds some risks to both CT and MRI, but it is also often necessary for the interpretation of images, explained Dr. Kasprzak. “Each contrast medium administration should be considered a risk-benefit analysis,” he said.

Risks include allergies, which are more common with CT than MRI. The nature of the allergy and of the current condition should drive decision making about whether and when to scan (for example, after preparing for an allergic reaction with medication).

“A patient who sneezed or had some hives after CT contrast is much less of an issue than a patient who had angioedema,” he said. “If you have a patient that you're worried about acute dissection … I don't know if you're going to want to wait for the prep.” Newer contrast agents, which are more osmotic and nonionic, are less likely to cause reactions, he said.

Contrast-induced nephropathy has been another concern with CT scans, but it might not be the common risk it is traditionally believed to be, according to Dr. Kasprzak. “Some of the more recent data that's come out basically suggests that potentially this is much, much less of a risk than we thought,” he said.

On the other hand, another risk has become the new focus of concern: the possibility that contrast exacerbates DNA damage caused by radiation. “This is something that we're not sure of the clinical significance of, whether or not this is tied to any radiation-induced sarcomas. This is something that's going to be popping up in the literature more, just as we're kind of resolving the contrast-induced nephropathy issue,” said Dr. Kasprzak.

Another new concern is associated with gadolinium, the MRI contrast agent that's already been identified as the cause of nephrogenic systemic fibrosis. Now it's been found to deposit itself in the basal ganglia after a scan. “We're not sure yet what, if any, clinical implication there is,” he said. The good news is that the newer macrocyclic MRI contrast agents shouldn't cause either problem.

All these risks can scare physicians and patients off contrast, but it is a very valuable tool in radiology, Dr. Kasprzak noted. “There are things that we as radiologists can do with noncontrast studies, but contrast really, really increases the conspicuity of findings,” he said. “I like to be definitive, but I can't do that without contrast. When possible, when the ACR appropriateness criteria advocates for it, let's use contrast.”

In addition to the potential for accurate diagnosis, use of the criteria is about to be financially encouraged. “CMS has established guidelines about appropriate use criteria,” he said. “What'll be happening in the future is there will be a software program superimposed on your [electronic medical record] and basically if you go to order a foot MRI to exclude [pulmonary embolism], the software is going to say, ‘No, you can't do that.’”

That “no” will either be a hard stop, which absolutely prevents the ordering of the test, or a soft stop, which can be overridden with a justification, but the involvement of CMS will make it difficult to order any test that isn't judged appropriate, predicted Dr. Kasprzak. “The hospital is not going to get paid, so they're not going to allow you to order these studies,” he said.

That's only the first step in computers guiding radiology practice, Dr. Kasprzak predicted. “On the diagnostic end, artificial intelligence or big data is going to make a big impact when it comes to pattern recognition,” he said. “I think the role of radiologist will go back to what I envision it should be, which is truly a consultant. I may be rounding with you or doing studies with you, doing things in a very different way than you're used to having a radiologist interact.”

Aside from the obvious benefits to practice, this new setup also will make it harder for other physicians to mock radiologists behind their backs.