July 2010

Reducing and Tracking Dose
By David Yeager
Radiology Today
Vol. 11 No. 7 P. 14

Cutting radiation exposure per scan is only part of the safety equation. Recording, reporting, tracking dose, and building cumulative exposure information into medical records is the ultimate objective. The work toward that goal is just starting.

Reducing CT scan dose is a hot topic these days, and it’s not surprising after several disturbing news reports last year. More than 200 patients undergoing CT brain perfusion scans at one hospital received approximately eight times the necessary radiation dose. A 2-year-old boy received 151 head scans during one exam. The New York Times ran a series of stories on radiation exposure, including one about a man who received fatal doses of radiation therapy. Even Congress has taken an interest in dose reduction. On February 26, the Subcommittee on Health of the House Committee on Energy and Commerce held a hearing to examine the issue.

The concern is understandable. A typical chest CT produces approximately 50 times the amount of radiation of a chest x-ray so cutting dose can have a significant effect on immediate and lifetime radiation exposure. Radiation exposure is a relatively new topic to the general public and it has turned up the heat in the radiology and the imaging industry, where it has always been a concern.

“Most vendors can show 10 years of dose-reduction technology poured into the product,” says Dominic Smith, vice president of marketing for CT and nuclear medicine for Philips Healthcare. “So it’s not like we woke up one day and figured that dose reduction was a good thing. And, for most of us, it’s our family members, it’s our loved ones, too. We know that this is a great technology, but we know that we have to reduce the natural side effect, as it were.”

In the last few years, each of the major CT manufacturers has released a large, very fast scanner. GE has Discovery CT 750 HD, Philips has Brilliance iCT, Siemens has Somatom Definition Flash, and Toshiba has Aquilion ONE. The speed at which they obtain a scan is so fast that a patient’s radiation exposure is cut dramatically.

In addition, new technologies have been added to scanners that help reduce dose. All the major manufacturers feature some version of these. One method is active collimation, where the collimator blades open and close independently of each other. Prior to active collimation, it was necessary to overrange past the area of interest. With single-slice scanners, this wasn’t much of a problem. But when CT machines grew to 64 slices, the overrange areas became significantly larger.

“What that does is cut off the unneeded x-rays in the overranged part and we can reduce the dose anywhere from 5% to about 30% for helical scanning,” says Rich Mather, senior manager of clinical programs at Toshiba Medical Research Institute. “And that depends on the width of your detector, and it depends on the length of your scan.”

Another method is dose modulation, which varies the amount of dose based on a patient’s shape and size. The system takes a picture of the body and forms a computer model of it. The more dense parts of the body receive a higher radiation dose while parts that are less dense receive a smaller amount.

“You acquire the topogram—which looks like a normal x-ray and provides us with attenuation information for the patient,” says Christian Eusemann, scientific collaboration manager of CT for Siemens Healthcare. “And, for example, if we know that we are looking in the abdomen, where there is more attenuation than in the chest, the system automatically increases the dose, the mA [milliamperes] in that case.” Eusemann says this technique can reduce dose by nearly 70%.

New Currents
The dose-reduction technique that’s currently generating the most buzz is iterative reconstruction. According to those in attendance, it was frequently discussed at the International Society for Computed Tomography’s symposium at Stanford University in May. Each of the large CT vendors has a version of this technology: GE has ASIR (Adaptive Statistical Iterative Reconstruction), Philips has iDose, Siemens has IRIS (Iterative Reconstruction in Image Space), and Toshiba has AIDR (Adaptive Iterative Dose Reduction).

Rather than taking one look at the data, as in standard filtered back projection, the reconstruction engine makes several passes over the data to produce a more accurate model of the image and reduce the amount of noise. These extra computations also allow the image to be acquired with a much lower dose, anywhere from 40% to 80% less depending on the manufacturer, the type of scan, and the scanner that’s used. The trade-off is that it requires more processing power and adds time to the process. However, the high quality of the images has led manufacturers to look beyond adaptive iterative reconstruction to full, model-based iterative reconstruction, even though it requires significantly more processing power and takes longer to process.

“Now what this does is it takes some of the technologies that we’ve put in play with our adaptive statistical iterative reconstruction, which models the noise in the system and models some of the physics, and does a full model-based approach,” says Eric Stahre, general manager of premium CT for GE Healthcare. “So it models the imaging chain, the whole imaging physics behind it, as well as the noise. And this is technology that’s currently in development.”

While all of these technologies can contribute to a significant drop in patients’ radiation exposure, one aspect of dose reduction sometimes overlooked is user education. “One great approach to dose reduction is, really, education of technologists who use the system, making sure that they understand all the tools and are putting them to the best use,” says Mather. “Customers who have invested in the education of their staff are able to maximize the capabilities of the technology, including dose reduction.”

CT Dose Check Initiative
In addition to better education, the inclusion of fail-safe mechanisms on scanners can help users prevent overdosing. In testimony before the Subcommittee on Health in February, David Fisher, a vice president at the National Electrical Manufacturers Association and managing director of the Medical Imaging & Technology Alliance, of which all the large CT manufacturers are members, introduced the CT Dose Check Initiative.

The first of the initiative’s three parts is a dose notification that lets the operator know whether a particular scan’s dose is higher than a reference-dose level. This feature is intended to reduce cumulative dose over time.

The second part is a dose alert that lets the user know whether there is too much radiation associated with a particular scan. This feature is intended to prevent overdosing on a particular scan.

The third part is a dose-recording standard. CT machines currently record dose data on the actual CT images but, since they travel with the images, they are difficult to track and analyze. To address this problem, the major vendors have all committed to using DICOM Structured Reporting (SR) by the end of the year to track dose. The use of DICOM SR will allow dose information to be recorded more easily by making it searchable and more easily incorporated into patient records.

“What we have now with the dose-structured reporting is that, we are actually writing out dose information into a document you can save as a PDF, an HTML, or an XML if you want, and that allows you to better analyze the data,” says Eusemann. “So you can actually track dose information in more detail. This is a major step for integrating putting dose information into a patient record and allows us to work with national dose registries and so on.”

Tracking dose is an essential step in helping patients estimate their lifetime exposure to radiation. However, there has been little emphasis on dose tracking up to now. In her testimony at February’s Subcommittee on Health hearing, Rebecca Smith-Bindman, MD, lead author of the study “Radiation Dose Associated With Common Computed Tomography Examinations and the Associated Lifetime Attributable Risk of Cancer,” pointed out that efforts to track dose have barely scratched the surface.

“Since information on radiation is reported differently across the different types of CT machines, it is difficult for radiologists to standardize their practice,” Smith-Bindman said. “The American College of Radiology has established a voluntary accreditation process to try to standardize practice and collects dose information but only on a very small sample of tests. This approach is promising, but at this point in time, the data collection is extremely limited, making it difficult if not impossible for the college to monitor if facilities comply with their recommendations.”

Part of the problem is that there hasn’t been much demand for the capability. “The demand for this in the marketplace, if you will, wasn’t there until very recently, and so I think we’ll see a bit of a scramble on the part of IT vendors to come up with a potential solution,” says Ken Denison, CT dose leader for GE Healthcare. “We’ve been approached by several large healthcare providers to help them think through this, and we’re looking at potential pilot projects but, at this point, that’s where that stands.”

“I think the digital information exists and as we become more consumers of health, you’ll be following your own personal health agenda, as it were,” says Smith. “We definitely see this as a trend.”

Ride of the Valkyrie
Now that demand is growing, manufacturers will work to make the data available—but that’s only half of the equation. “Our job as manufacturers is to record the information,” says Fisher. “We’re not really reporting; we’re recording.” It will be up to healthcare organizations to figure out how to make that information available to patients. A new system called Valkyrie is the first attempt to bridge that gap.

Valkyrie is an automated stand-alone system created through a collaboration between Weill Cornell Medical Center and Columbia University Medical Center in New York. Once a study is done, it automatically flows to the system. In 15 to 20 seconds, the dose information is extracted. A patient’s information is automatically sent to the system, and the Valkyrie adjusts for patient positioning, size, age, and gender. This provides a highly accurate estimate of what the patient’s absorbed dose is rather than the dose that’s emitted by the CT scanner. None of the major vendors has this capability yet.

Although Valkyrie was initially developed to help with quality control, George Shih, MD, an assistant professor and the director of imaging informatics in the radiology department at NewYork-Presbyterian Hospital/Weill Cornell Medical Center, believes it could be used to populate an EHR or a PHR.

Valkyrie was mainly tested in conjunction with GE scanners and, to a lesser degree, Siemens scanners. Shih is confident that it will work with any scanner though. And while he doesn’t see it as a long-term solution, he believes it fills an important niche.

“I kind of think of Valkyrie as more of an immediate solution to try to get the dose information that you should have been easily available years ago,” says Shih. “I don’t see it being a great solution in 10 or 20 years or anything like that, but I don’t think we need to wait years to start collecting this data, which may have both clinical and research benefits.

“And, certainly, the one advantage that Valkyrie has over some of the initiatives, such as DICOM SR, is that we can go back and look retrospectively at the studies in any PACS that are already done and archived and calculate those studies from existing data,” he adds.

Although one of Shih’s main goals is to make dose information part of patients’ medical records as soon as possible, he says he’s received calls from large and small vendors about collaborating with Valkyrie.

“While our primary goal was really patient safety at our institution, we would be happy if institutions started using Valkyrie or something like Valkyrie,” he says. “Although there may be other good solutions out there, we’re happy to work with anyone who’s interested. I think this is really important as a field so the more we can help each other out, the better because I think there are probably some great opportunities to further minimize CT radiation dose and inadvertent radiation overexposure events.”

— David Yeager is a freelance writer and editor based in Royersford, Pa.