Public scrutiny recently has intensified toward physicians’ use of CT (computed tomography) to help diagnose diseases and disorders. At issue is patients’ potentially unnecessary exposure to radiation. Dr. William Shuman is vice-chair of UW’s Department of Radiology. He discussed use of CT at UW Medical Center, where he has been radiology director since 2003.
Q. When did the medical and lay communities start calling CT use overzealous?
A. Three to four years ago we started hearing concerns: The growth of CT was double-digit over about an eight-year period, far greater than the population growth. The primary driver for that was the central role CT came to play in diagnosis to get proper therapy going. There were economic drivers too, though I don’t think they were prevalent in the Northwest. Since then, the focus on CT value and appropriateness has expanded rapidly.
Q. Is it fair to say that some CT is used inappropriately?
A. Yes. The number you hear quoted all the time in news reports is that 30 percent of scans may be unnecessary or are not totally indicated when they are ordered, but I can find no data to back up that number. Here at UW Medicine, I’d be surprised if 2 percent of our studies would fall into an “inappropriate” category by national standards established by either data-driven literature or opinion published by the American or European College of Radiology.
When you use those kinds of criteria, which I think the industry should, both from the perspective of payers and providers, then I think we are extremely low.
Q. How does UW Medicine verify that people derive benefit from CT?
A. Remember, there are two concerns here: one is minimizing radiation dose, the other is stopping inappropriate exams. In the latter category, we screen seven times for appropriateness before a patient gets on a scanner. Other actions have been largely IT-driven: We do electronic protocoling now.
Q. What is that?
A. We have a selection of 800 or 900 protocols that we can do with CT, one of which is going to be appropriate for that patient with that indication. There are dozens of variables to consider, for instance, parameters of the scanner, the contrast we inject, whether the patient ingests contrast ahead of time, and so on. It’s extremely detailed but all choices are pre-existing.
We sit down with our list of patients who will be scanned, and for each patient we go over the electronic medical record, the requisition from the referring physician, the lab values, allergy record, kidney-function tests. Based on all that, we decide the protocol for each patient that we think will give the most information possible with the least amount of radiation. We electronically select that, can add modifications, and then we approve it for the technologist who will execute the study.
We also put huge effort into training referring clinicians about appropriate use. We will be implementing electronic clinical-decision support – software that operates in the electronic medical record and requires the referring physician to answer five quick questions. It generates an immediate judgment based on more than 1,000 published criteria whether CT is appropriate, of questionable merit, or inappropriate.
Q. What about minimizing radiation dosage?
A. There’s standard dose, which is widely available in the community; then there’s low dose, which is standard minus 40 percent of the radiation, and ultralow dose, which is standard minus 90 percent.
Maybe one-fifth of the scanners in WWAMI can do low dose. Ultralow requires a software change and bank of about 15 servers; it’s a fairly expensive upgrade and we have, to my knowledge, the only one in the Pacific Northwest. It’s a piece of technology, but there’s the technique, as well. You need both to get to ultralow dose. Any scanner more than about two years old will not be able to do low dose or ultralow dose.
New model scanners (just approved by the FDA) that we just installed in the Montlake Tower enable us to offer ultralow-dose protocols. These produce the lowest-dose CT scans west of the Mississippi.
Q. It sounds like if you don’t have software and hardware made in the past two to three years, you’re outmoded.
A. That’s true. What’s gotten better is the reconstruction of raw data into an image. Think of a satellite 100 miles above the Earth and taking a picture of a parking lot and being able to read license plate numbers. This takes software crunching raw data, which has evolved rapidly in the past few years.
You get the technology to do that, then the question becomes, how far in our local situation can we push the technology? For example, how low can we take the dose before we start to affect our diagnostic power? Which patients should get ultralow dose and who doesn’t need it? In what studies does it have the greatest impact? We investigate those questions at academic medical centers.
Q. Why wouldn’t every patient need or want ultralow dose?
A. The time to go from scanning to images available is one to two minutes with low-dose scanning. For ultralow-dose, it’s over an hour because so many complex calculations are involved. So it’s not for everybody.
Q. How good is communication between doctors who treat one person in multiple service areas, say gastroenterology and head-neck? Do patients fall through the cracks and get too much radiation?
A. We have all that information in the radiology information system and the inpatient electronic medical record. Usually patients bring us their CD of scans from other providers, and it goes straight into our PACS (Picture Archiving and Communications System) here. So everyone here is aware of what the patient knows.
Q. Is it accurate to say that standard CT scans are equal to 350 x-rays?
A. If you use the chest X-ray as comparison, which is the lowest-dose imaging we do, then it is equivalent. But patients have no idea the amount of radiation involved in a chest X-ray. We try to explain it in terms of risks that people can understand. In a CT exam of the abdomen and pelvis with standard technique, patients are exposed to a cancer risk about equal to the chance of dying in a car accident if you drive a car 70,000 miles over four years in the U.S.
What we really are concerned about is not risk, alone, but risk-benefit ratio. The benefit from CT, if used appropriately, can very often be thousands of times in a patient’s favor. CT does so many things well, looks at the whole body, does angiograms, helps guide biopsies. It’s certainly the most highly regarded tool in the toolbox in terms of its impact on the practice of medicine. I mean, there’s a reason Dr. Godfrey Hounsfield got the Nobel Prize.
To refer a patient or learn more, call 206.598.7200.