|
Advanced CT is making it possible to image faster, with thinner slices and greater range than ever before. It is also reshaping the way contrast material is delivered, prompting radiologists to reexamine everything from delivery rates to contrast volumes to iodine concentrations.
"With 16-slice, you have to rethink how you do CT. There's no way around it," said Dr. Elliot K. Fishman, director of diagnostic radiology and body CT at Johns Hopkins Hospital.
A scan that takes less than 10 seconds leaves little room for error. Imaging at peak contrast enhancement requires not only precise timing but careful tailoring for each patient.
"It's crucial that you revise your protocols. The quality of the images you interpret is only as good as the time you spend making sure your protocols are correct for the disease process you're trying to evaluate," said Dr. Paul H. Silverman, who holds the Gerald D. Dodd Jr. Distinguished Chair in Diagnostic Imaging at the University of Texas M.D. Anderson Cancer Center.
Higher contrast injection rates were among the first changes prompted by multidetector imaging. In 1995 and 1998, Silverman published the results of two surveys he conducted among members of the Society of Computed Body Tomography and Magnetic Resonance. The surveys showed that even following the introduction of four-slice scanners, injection rates for routine body examinations increased to 2.5 to 3 mL/sec, from a starting point of about 1 to 1.5 mL/sec.
These data reflect a major change in priorities, from prolonging contrast administration with conventional CT to delivering contrast fast enough to achieve peak enhancement during rapid image acquisition.
"People don't drip contrast anymore. They don't infuse contrast anymore. They inject a bolus of contrast with a power injector," Silverman said.
Typical injection rates are even higher with vascular imaging. Fishman reports using injection rates of 3.5 to 4 mL/sec for vascular studies. Some radiologists, particularly in Europe, are pushing rates even higher, to 7 or 8 mL/sec. Such rates are not common in the U.S., however.
"There's too much risk of complication," Fishman said. "You need large-bore catheters, and many radiologists are uncomfortable with them, as they fear the risk of extravasation. And the gain in image quality and diagnostic information may not be there."
While fast contrast injection clearly improves vascular imaging, determining ideal injection rates for parenchymal imaging is more complex.
"It depends on the organ's vascular supply and what you're trying to identify," said Dr. Jay Heiken, director of abdominal imaging at Washington University in St. Louis.
The pancreas, for example, is supplied primarily through arterial flow. Rapid injection of contrast material therefore improves pancreatic enhancement, just as in vascular imaging. By comparison, the liver receives the majority of its blood supply through the portal vein. Heiken's research has shown that injection rates exceeding 2 to 3 mL/sec only minimally improve hepatic parenchymal enhancement. One exception is in the identification of hypervascular liver lesions, such as those associated with hepatocellular carcinoma. In that case, higher injection rates are helpful, as imaging takes place during the arterial phase of enhancement.
In Silverman's surveys, injection rates for the evaluation of hypervascular lesions averaged 5 to 6 mL/sec.
TIMING
The ability to time image acquisition to coincide with peak contrast enhancement is both the promise and the challenge of fast CT scanning. The risk is not in scanning too late, as in the days of conventional CT, but in scanning too early.
"Now, because of the speed, you have to be very careful that you're scanning when there's enough contrast in place," Fishman said. "If you're not careful, you can still be injecting when the scan is over."
Timing is particularly critical in the liver. Most liver lesions are hypovascular, taking up contrast more slowly than liver parenchyma, so the better the enhancement of the liver, the better the conspicuity of the comparatively dark lesion. The ideal time to image hypovascular liver lesions is 60 to 70 seconds after injection, during the peak of the portal venous phase.
"But over just tens of seconds, or sometimes even seconds, contrast diffuses into the interstitial spaces," Silverman said. "If you scan the lesion even a half-minute later, you have much poorer detectability because equilibration allows the abnormal tissue to enhance."
Hypervascular lesions, on the other hand, enhance before the liver does. As a result, hypervascular lesions appear bright on the background of a dark liver when scanning takes place just 20 to 30 seconds after contrast injection, during the hepatic arterial phase of enhancement. Hypervascular lesions include not only hepatocellular cancer, but also metastases from endocrine tumors, renal cell carcinoma, carcinoid tumors, melanoma, sarcoma, and choriocarcinoma.
Precise timing of image acquisition can be tricky. Contrast transit times vary from patient to patient, depending on body size, cardiac output, clinical condition, and normal physiologic variation. Silverman favors computer-automated scanning technology, also called bolus-tracking or automated triggering, for determining the optimal delay between injection and scanning. Using this technique, a series of low-radiation scans track the progress of the contrast bolus. Cursors can be placed on a region of interest and the scanner set to begin image acquisition as soon as enhancement reaches a predetermined threshold.
"This allows radiologists to adapt the scan delay to each patient's circulation time," he said.
Automated triggering does not have universal support, however. A common complaint is that the time between the arrival of the contrast bolus and the initiation of scanning can be too long with automated programs. This problem varies among manufacturers and is improving, according to several sources. In addition, automated triggering may be less effective for peripheral vascular studies, since circulation times in the two legs often differ in patients with vascular disease. Such programs also demand more of radiology technologists.
For the last reason, Heiken routinely uses bolus tracking for vascular imaging but not for routine examinations of the chest or abdomen, where timing isn't as critical. Instead, he relies on preset scan delays for those examinations.
Fishman doesn't use automated triggering at all, not even for vascular imaging. Not only is it possible to achieve excellent image quality with preset delays, he said, but this approach enables him to ensure the safety of patients and technologists.
"I'm more worried about the extravasation than being a little bit off in the timing," he said.
To monitor patients for extravasation, Johns Hopkins technologists remain with the patient while the first 50 to 75 mL of contrast is injected, manually palpating the injection site to ensure that the intravenous line is properly placed and flowing smoothly. Before the scan begins, they move behind the lead-shielded wall. With bolus tracking, a technologist would be repeatedly exposed to low-level radiation while palpating the IV site, a potential problem for women of childbearing age.
"You could do a test injection to make sure the IV is in and not bother staying there, but we have found you can get extravasation even though your test bolus works well," Fishman said.
CONCENTRATION
As scan times shrink, radiologists are improving the density of enhancement not only through faster contrast injection but also through use of higher concentration contrast material. In the days of conventional CT, iodine concentrations rarely exceeded 300 mg/mL, Silverman said. With the high volumes necessitated by prolonged scans, higher concentration contrast material would have resulted in the delivery of a toxic iodine load.
Today, the use of 350 mg/mL of iodine contrast material is typical. According to Heiken, its primary benefit is in vascular imaging. Injecting a higher concentration contrast agent in a shorter period of time during CT angiography clearly increases vascular enhancement. Its benefits in parenchymal imaging are less apparent.
Researchers are pushing ahead, investigating the use of contrast material with an iodine concentration of 400 mg/mL. Whether it will prove clinically useful and practical is unclear. One problem is that increasing the iodine concentration also increases viscosity, which makes it more difficult to inject at high flow rates. More fundamentally, higher concentration contrast material may be unnecessary for obtaining outstanding images.
"If you do volume rendering or MIPs very well, you don't need the higher concentration," Fishman said. "You can use the computer to optimize the quality of the 3D images."
Still, with every healthcare dollar under scrutiny, higher concentration contrast agents may offer a hidden benefit: cost-savings.
"Since price is generally dependent on volume, you can give the same number of grams of iodine in a more concentrated solution and decrease the cost," Silverman said.
VOLUME
The ability to reduce contrast volume is an appealing bonus of shorter acquisition times. The degree to which contrast volume can be decreased depends on the study, however. Many radiologists have been able to cut contrast volume by roughly half, to no more than 80 mL, for routine examinations of the chest or aortoiliac vessels, among others.
"With less contrast, you're still getting an excellent examination," Silverman said.
Even vascular studies that extend from the abdominal aorta through the feet can be done with 120 mL of contrast, a reduction of 25% to 40% from the volumes typically used with single- and four-slice scanning.
Whether, and how much, contrast volume can be cut during liver imaging is controversial. A certain amount of iodine is needed to achieve adequate parenchymal enhancement, and it's not possible to go below that amount without sacrificing image quality. Heiken's research has suggested that 35 to 40 g of iodine is the minimum, which roughly equates to 100 to 115 mL of 350 mg/mL contrast material. (Heiken actually uses 125 mL of contrast.)
"There were others who disagreed, but my interpretation of the studies was that we could reduce the total iodine load by about 20% because we were scanning during a shorter period of time and could concentrate our imaging during the peak of parenchymal enhancement," he said. "That's even more true with the newer scanners because of improvements in spatial and temporal resolution."
SALINE FLUSH
With the development of double-barreled power injectors, the role of the saline flush is gaining new attention. Immediately following contrast with an injection of saline has several potential benefits; chief among them is the opportunity to take advantage of contrast material that lingers in the intravenous tubing and arm veins-easily 15 mL altogether. Use of the saline flush has the potential to prolong and increase enhancement by maintaining a "tighter" contrast bolus, and to reduce contrast volume.
"It's better for the patient," Fishman said. "If you can get the same study with less contrast, why not do it?"
Not many radiologists have embraced the saline flush, however. Dr. Giovanna Casola, director of body imaging at the University of California, San Diego, said that although the saline flush has theoretical benefits, today's delivery systems need improvement. Even in the case of CT urography, which involves infusion of additional fluid to distend the collecting system of the kidney, saline injection is inferior to simpler methods.
"We're not using the saline flush injector. For our CT urography, we're simply hooking up a bag of saline to the IV after the injection," Casola said.
Faster scanners may even outpace the saline flush in some cases, according to Heiken. In chest imaging, for example, dense contrast material in the superior vena cava and axillary vein can create artifact. In theory, a saline flush should evacuate dense contrast from these vessels; in fact, it usually doesn't get the chance.
"We've found that when you're using a 16-slice scanner, sometimes the imaging is done so rapidly that there is not enough time to flush the contrast through," Heiken said. "We've been inconsistent in getting a good result."
Almost all radiologists agree on one fact: With advanced CT, use of contrast may be more common, but its delivery is far from routine.
"It requires much more tailoring," Casola said. "The physician input has to be there, and you have to pay much more attention to your protocols and techniques."
MS. CARRINGTON is a freelance medical writer in Vallejo, CA.
| 
