Advances in CT technology have generally met with unrestrained enthusiasm. In peripheral vascular imaging, however, some experts are beginning to question whether more really is better.
Four-slice scanners made peripheral CT angiography possible, and they have played a role in nearly all published research validating the technique. Experts readily point to the advantages of eight- and 16-slice CT scanners as well.
"An eight- or 16-detector scanner allows you to scan faster and with thinner slices than a four-detector scanner," said Dr. Michael Martin, an assistant professor of radiology at the University of British Columbia. "You have better spatial resolution and far less helical reconstruction, motion, and breathing artifact."
Martin uses four-, eight-, and 16-slice CT scanners in his busy Vancouver practice. Improved spatial resolution is the payoff everyone using such scanners talks about first.
"When we scanned the peripheral vasculature with the four-row scanner, we used a 2.5-mm thickness. Ever since we went to the eight-row scanner, we have used a 1.25-mm thickness," said Dr. Geoffrey Rubin, a professor of medicine and chief of cardiovascular imaging at Stanford University School of Medicine. "There is no question that, from a visual standpoint, image quality is substantially improved."
With the introduction of 32-, 40-, and 64-slice scanners, however, the response is more restrained. Potential improvements in spatial and temporal resolution must be balanced against increased image noise, concerns about radiation dose, and more complex contrast administration protocols. The trade-off leaves some radiologists skeptical about the role of the most advanced multislice CT scanners for imaging the lower extremities.
"If there are advantages with the 64-slice scanner for peripheral vascular imaging, they're incremental and relatively fine increments at that, compared with the advances we saw with the four-row scanner, and then the eight and 16," Rubin said.
Higher spatial resolution has the potential to improve not only the visual quality of images, but also diagnosis. With eight- and 16-slice scanners, it is possible to image lower in the legs with greater confidence, according to Martin.
"It's easier to interpret those images than with a four-slice scanner, unquestionably," he said.
Dr. Tobias Jakobs, a radiologist at the University of Munich, Grosshadern campus, has been working with a 64-slice scanner for several months. He notes that this latest advance in scanner technology adds further value to peripheral vascular imaging, but only in certain clinical situations. A patient with an acute onset of symptoms suggestive of embolic disease might not need imaging with an advanced scanner.
"If the patient has an acute onset, there might be occlusion of a major vessel, and you can easily detect this with even a four-detector-row CT, with a high confidence," Jakobs said.
Improved spatial resolution may pay off more in other cases. The 16-slice scanner enables detection of stenoses down to the popliteal arteries, but the 64-slice scanner makes it possible to more accurately grade stenoses and determine their hemodynamic significance, particularly in the case of a patient with multiple lesions, he said.
A 64-slice scanner might offer an advantage in a patient with chronic peripheral vascular disease (PVD). On follow-up examination, the scanner must be able to detect subtle changes in disease burden.
"It is really helpful to have the higher spatial resolution for the chronic disease patient," Jakobs said. "It can change management."
Although submillimeter slice widths are possible on all the latest scanners, many radiologists select a detector collimation of 1 to 1.5 mm when imaging a patient with PVD.
"Although our 16-slice scanner is capable of 0.625-mm slice thickness, we choose not to use the narrowest possible thickness," Martin said. "We found that it did not add to diagnostic accuracy. The images were noisier and grainier and much more difficult to interpret."
The potential drawbacks of submillimeter slice widths in peripheral vascular imaging are more obvious in some portions of the exam than others.
"Many regions of the body don't tolerate that much noise, particularly the aorta and its branches. When you're talking about peripheral arterial disease, you cannot consider the legs without considering the aorta and iliac arteries, because inflow is critical," Rubin said.
Although very thin slice widths are likely to improve image quality from the knees to the toes, the impact even in the thighs may vary, depending on the size of the thighs. And imaging with very thin sections in the abdomen could take careful planning.
"For many of our patients, the noise levels that result from using that very thin detector width outweigh the benefits of the added spatial resolution," Rubin said.
The interplay of speed, slice thickness, and radiation dose isn't the only complication in decisions dictating which scanner to choose. Other factors come into play.
-Blooming artifact. This inherent shortcoming of CTA can result in overestimation of the size of calcified lesions and underestimation of luminal diameter. With the 64-slice scanner, problems with blooming artifact are less common, Jakobs said.
"We do not have much blooming artifact anymore," he said. "I think with the 64-slice scanner, you can be sure what you see really is the lumen. You can be more confident in making the correct diagnosis."
-Scan range. A clear benefit of advanced scanner technology has been an improvement in scan range. Today, all relevant anatomy can be included in a distal runoff study.
"With the four-detector scanner, we're limited to imaging roughly from the level of the celiac artery to the ankles. When you go to an eight- or 16-detector scanner, there is essentially no limitation. You can go from the top of the chest to the tips of the toes without any problem," Martin said.
The 64-slice scanner would make it easy to go even further and scan from the neck to the toes. Such an approach would give a nod to the concept of atherosclerosis as a systemic disease and enable detection of plaque in the carotid and coronary arteries, thoracic and abdominal aorta, and lower extremities. This prospect appeals not just to radiologists but to other cardiovascular physicians as well.
"When I was in training, we were interested only in the heart, and all these other vessels were just a way to get to the heart. Over the years, we've come to the understanding that it is all part of the same disease process," said Dr. Peter S. Fail, director of cardiac catheterization and interventional research at the Cardiovascular Institute of the South in Lafayette, LA.
It may be a while before that vision is realized. In the absence of validation and outcomes studies, total-body CTA is not yet considered justified in everyday practice, although it is a focus of research at Stanford and other institutions.
"Scanning people from neck to toes to see where they have atherosclerosis and using that information to make treatment decisions could come in the future, but I think it's pretty far off," Rubin said.
-Temporal resolution. Gantry rotation times have become even faster on some of the most advanced CT scanners, ranging from 0.33 to 0.4 second. Although a reduction in temporal resolution is helpful in imaging the coronary arteries and other vessels prone to pulsation artifact, it has less impact when imaging the peripheral vasculature.
The combination of thinner slices and faster gantry rotation may necessitate scanning patients at a higher mA. This places greater demands on the x-ray tube and also increases the radiation dose. Several vendors have responded by redesigning or enhancing the x-ray tube in the latest generation of scanners. Dose-management software can modulate tube output according to the type of scan and region of the body. In this way, tube output is at its maximum only when necessary to obtain diagnostic-quality images, minimizing both radiation exposure and the risk that the x-ray tube will overheat.
The experts agree on the bottom line: CTA of the peripheral vasculature can be accomplished on a wide range of scanners, but some are better than others for particular uses.
"If you work in an institution with only a four-detector scanner, you should still be offering this as a modality for investigation of lower limb ischemia," Martin said. "But if you have a choice between a four and an eight, go for the eight."
Martin considers a four-slice scanner sufficient for imaging most patients with claudication. Better spatial resolution is necessary, however, when imaging patients with critical ischemia of the legs, rest pain, tissue loss, or ulceration. These conditions suggest a more significant degree of arterial insufficiency, often involving smaller vessels.
"When the contrast bolus is properly handled, we see vessels as small as any surgeon would ever target, using an eight- or 16-row scanner and a 1.25-mm thickness," he said. "We've scanned two-year-olds with crush injuries and have visualized 1-mm vessels in their feet. I don't think that we need to see vessels smaller than that."
Properly handling the contrast bolus becomes increasingly tricky as scan speeds rise. With the introduction of the 16-slice scanner, radiologists found it necessary to adjust contrast and image acquisition protocols or risk having the scanner outrun the contrast bolus.
"To some degree, you have to scan slower than the scanner is capable of," Martin said. "If you don't, your table will move faster than the blood flow, and by the time you get to the calves, there will be no more arterial enhancement."
Making this adjustment requires counterintuitive thinking. When scan times fall to 20 seconds or less, the natural inclination is to reduce the duration of contrast injection accordingly and increase the injection rate.
But research at Stanford suggests that injecting faster and for a shorter time is not the best approach to use when imaging patients with PVD. There can be a sixfold difference from one patient to another in the time it takes for contrast material to flow from the aorta to the popliteal arteries. And a patient's flow rate cannot be predicted by the severity of disease.
Therefore, both Martin and Rubin inject contrast material over about 35 seconds, regardless of the type of scanner they are using. Rubin uses automated bolus triggering and adds a 20-second delay in addition to the time to contrast arrival in the aorta.
"We start out behind, but not too far behind," Rubin said. "If the patient's flow rate is very slow, we catch up to the head of the bolus by the time we get to the feet. But if the flow rate is very fast, we still have enough bolus behind us that we don't lose out."
The rate of contrast injection is also determined by patient size. In a large patient, Rubin injects 350 to 370 mgI/mL contrast material at 5 mL/sec, whereas in a small patient, he injects contrast at 4 mL/sec (in both cases, over 35 seconds). Ultimately, a good contrast injection protocol can be as critical as scanner technology, he said.
"Anyone who isn't seeing distal vessels very well would be better served by reviewing their contrast delivery protocol than by getting the latest scanner," Rubin said. "The key to visualization of those small vessels is the concentration of iodine in that artery when the scanner arrives there."