Applied Radiology

Moderator: Stuart E. Mirvis, MD, FACR
Professor of Radiology in the Diagnostic Imaging Department, University of Maryland Medical Center, Baltimore, MD and Editor-in-Chief of Applied Radiology

STUART MIRVIS, MD, FACR: From our panelists' presentations, it is clear that there are challenges in using fast CT and in optimizing contrast dose. I'd like to focus our discussion on contrast enhancement and its use in some of the body systems.

I want to address the advantages or disadvantages of using 370 mgI/mL concentration contrast material and some broader issues. Is that concentration adequate for all of our needs? Is there any reason to go higher than that? If a higher iodine concentration became available, would we want to use that product in our practices? Could this higher concentration iodine be useful to us in ways besides just enhancement? Could we benefit the patient directly? Does it help the institution in any way?

Finally, what is the best form for that contrast material to take? Are we comfortable in using prepackaged dosing? Do we want giant drums of contrast that we can pull from? What's the best way for our technicians to administer the material?

Let me start with a question to Dr. Foley. Dennis, you made a lot of excellent points about contrast scanning and spiral CT. What I didn't get from your lecture is which contrast you're using currently and how a higher strength iodine contrast might affect the results you have and how it might change your protocols.

DENNIS FOLEY, MD: We've been using 300 mgI/mL routinely for the last 3 to 4 years. All of our protocols have been based on that. It's fairly obvious, though, that as we go up in channel count and as we progress to isotropic voxels that we will benefit by having higher intra-arterial iodine concentration. So my expectation is that we will change our protocols.

In changing the approach to contrast injection, one can do either one of two things. You can take a 370 mgI/mL approach and have the same injection interval but lower the injection rate. Or, you can do the opposite and maintain the same injection rate and decrease the injection volume, provided that the CT scan is acquisitioned so we are able to acquire the data during the first circulation. In that particular scenario, you've also got to consider not only how fast the scanner can go, but what the image-quality tradeoff is of going extremely fast. My expectation is that we will probably maintain the same injection rate, decrease the volume, and still be able to acquire very good data.

Another thing to consider is the use of saline chasing. I think that as we potentially drop to lower volumes of injection contrast at a higher concentration, we're going to benefit by using saline-chase contrast injectors with the double-barrel syringe approach. It's common in MRI because with MRI, injection volumes are relatively small. To use those injection volumes efficiently, saline chase is used routinely. I think that companies that produce injectors for CT are following suit, as they see this trend toward lower volumes and CT angiography, to produce double-barreled injectors with the saline chase. So my expectation is that saline chasing and lower volume/higher concentration will probably go together.

SM: So if we go to lower volumes with higher concentrations and want to take advantage of all of the iodine available, we'll have to be sure to get that tight bolus where we need it when we need it. So this almost mandates some kind of chase. We're looking toward a whole new generation of injectors for CT.

DF: I think that's going to come, and I think it is a relatively small component of the total cost for institutions purchasing multidetector scanners. If it will improve the quality of the study, I think it should be acquired.

SM: Dr. Silverman, you're a strong advocate of high-concentration iodine contrast material and have been speaking about this for several years. What is your data on the higher iodine concentrations?

PAUL SILVERMAN, MD: In 1995, I realized that with the evolution of multislice scanners, we would soon be doing faster and faster scans. It's come to pass with the recent multi-slice CT scanners that we really need to go to a higher concentration of contrast to take advantage of the faster speed. That is simply because one does not need the volume anymore, so why should you use it? When we had slower scanners, we needed the volume.

Another increasingly important aspect is the need to become a very efficient CT operation. We've done that at M.D. Anderson using Six Sigma methodology, which has been shown to greatly improve efficiency. In addressing efficiency of operation, we can look at things like the use of prefilled syringes. Do they create a more efficient environment for you? Now, the answer can only be given that, yes, you will save time.

But that exact time is not going to mean the same for each individual, for each practice. If you take a significant amount of time out of the patient's scan for your technologist to load a syringe, then using prefilled syringes will have a tremendous impact. You may increase your throughput, for a day, for a week, for a month, for a year.

Another aspect that can increase efficiency or decrease cost is what Dr. Foley mentioned, a saline chase. You actually get a somewhat better contrast enhancement for the given amount of volume of contrast you purchase, so you get more for your money. The other way you can do it is by using the various computer-automated scanning technologies (CASTs). Each individual manufacturer has its own, with its own name. I think these are all aspects of healthcare efficiency that we have to look at very closely in a market that is certainly being constrained.

LARRY TANENBAUM, MD: We have found that by going to a higher concentration in a lower volume, in many circumstances we have been able to reduce the total amount of iodine.

Our institution has a reputation for being very efficient. We schedule 15-minute studies on our 8-channel CT scanner; that does challenge our facility's ability to deliver. A prefilled syringe might help us glean some significant time savings. We have the same considerations in MR. There are circumstances in which the technologists are busy and that 3 minutes they could save by using a prefilled syringe might be helpful.

The last point I want to make is about the saline chasers. Certainly Dr. Foley alluded to the fact that as contrast volumes continue to drop, the saline chase will become more and more valuable. But with the volumes we're using today, 100 or 50 mL for CTA, I don't know that there's a compelling need to throw away my new injector and buy one with a saline chaser. I'd hate to send that message with today's volumes. But when we move to higher concentration agents and reduce our volumes further, this is an area we may all get into.

RENDON NELSON, MD: What we don't know is whether the benefits of a saline chase are purely clearing out the tubing. Having saline at the back end of the bolus rather than blood may actually have some diagnostic advantages as well. I don't know.

DF: I'd just like to comment about SmartPrep and similar techniques. They are very useful for visceral imaging. But I think for angiography, particularly with the small boluses, timing is critical. We were still relying on having a mini-bolus to determine the patient's circulation time. Do you think that SmartPrep or technologies like it will be able to give you instant read-out?

PS: I think what has happened is that the drive for multislice technology, going from a 4-slice to an 8-slice to a 16-slice and to volumetric scanning in the future, has driven companies to neglect some of the other peripheral areas in CT. They're driven by the competition to have the fastest scanner, so some of this other technology has been neglected. I hope that in the future the companies will invest in making smart CT much smarter. It must be in order to do vascular studies, as Dr. Foley pointed out, because you just can't turn these around in less than about 6 or 8 seconds, therefore they're not that valuable on the arterial side.

RN: You can turn them around in 3 seconds. We use SmartPrep for all of our CT angiography.

PS: You can do that if you push it.

RN: Well, the 15 to 20 mL you use for your timing bolus, you said that you lose that if you have a 3-second delay. It's a lot more efficient to not have to do that.

PS: I think it's very interesting to see. Here you have two experts in the field, one does it one way and one does it another way. There's no clear-cut way to do it. The downside of a initial bolus or test bolus is you're wasting some contrast. The downside of some of the automated timing technology is that you may not be able to turn it around quite as quickly. You can do it in 3 or 4 seconds with a well-trained group of technologists. But I think Dr. Nelson has worked very hard with the technologists to train them to do that. So it really depends on the environment you're involved in.

RN: Although, I like having the collecting system opacified. That's nice.

LT : We are using SmartPrep for vascular imaging. And we're generically using an enhancement triggered scanning scheme. The latest generation of enhancement-triggered scanning is considerably more agile than the initial generation. The initial generation was not really practical for vascular imaging. As Dr. Nelson said, it's about a 3-second latency. All you need to do is be clever about the location you're scanning and you can pretty much use it as the trigger.

Getting back to efficiency, there's no separate injection, no separate calculation. There are some applications that are very challenging. Pulmonary angiography is a bit of a challenge. For coronary CTA, we're still timing boluses. But for most of our CTA, we're doing enhancement-triggered scanning, which is totally driven by the technologists. That's a very important feature; they just watch the heart enhance. As soon as they see the first flash, they start their scan.

Our success is close to 100%. The only time we've gotten into trouble is with people with very sluggish cardiac output. You see the contrast in the heart and assume it will then be at the renal arteries 3 seconds later and, surprise, it hasn't gotten there. But even that's predictable with training, by watching the way the contrast moves through the right heart. It is critical in terms of making this process run without direct radiologist supervision. At our institution, when we do CT from head-to-toe we don't need to supervise the performance of the procedure directly.

RN: I don't want to do a timing bolus. I don't want to calculate my dose. So we set the scanner at the top slice and start the power injector. As soon as the aorta enhances, we use automated triggering. We start the scan and simultaneously interrupt the power injector. Everything is interactive assuming that there is enough contrast upstream. We found the average dose and compared it randomly with the patients getting 150 mL. We got better enhancement uniformly through the aorta at 104 mL. I just cut it to 100 because you don't save any money by loading 104 mL. But we have better enhancement with less dose just by using that timing method.

SM: So you can set your targeting proximal to where you're going to be imaging?

RN: We set it right at that same spot. It's all qualitative. The tech just looks where it enhanced and they go.

LT: I was a little surprised to hear Dr. Foley be so meticulous about his techniques in terms of how long to inject and then how long to scan. I'm sure this is the product of some very careful research. Whereas, in clinical practice, we are often working with new machines before the research is done. We tend to fly a little bit more by the seat of our pants and do what seems to be correct until the literature catches up and proves one point or the other. But I find that CT angiography gives you a fairly broad window of opportunity. It isn't as much of a first-pass exam as it would seem to be ordinarily. So the fact that Dr. Nelson's sampling at the point of attack is 3 seconds difference, say, from sampling a few seconds proximal through the point of attack. I think you have a fairly broad window of opportunity with the fast contrast resolution of a CT angiogram.

SM: Effective and elegant.

DF: I'd just like to ask Dr. Nelson one question, though. If you are doing that, are you compensating by having your injection interval longer than your acquisition interval?

For example, if you have to scan the thoraco-aorta and pelvis in 20 seconds and you've got to trigger the aortic arch, how long is your injection interval for that type of study? Or does it vary with a patient's circulation time? You're saying you'd cut the injection off when you see the contrast in the aorta? So, in other words, your injection duration is variable?

RN: It is variable.

DF: Do you have a range for that?

RN: I don't remember what the range was for that. We probably use more than we need.

WILLIAM STANFORD, MD: We do a lot of our heart imaging with the electron beam. When we needed to give a continuous contrast injection to look at cardiac function, we got ready to stop after 3 seconds. Then, we turned the contrast off and watched the contrast bolus run through. But we knew that we would be through with our imaging by the end of the 12 seconds it took to get the tail end of a contrast bolus. Initially we ran the contrast the whole time we were doing our window. More recently, we've been starting contrast and then cutting it off because we know that it's going to take another 6 or 8 seconds for the tail end of that contrast bolus to get where we want it to be.

VAHID YAGHMAI, MD: In our experience, the number of diagnostic studies are much higher when you're using SmartPrep or you're using a time bolus. That probably has to do with the fact that you deal with a lot of older patients who have cardiovascular problems and once you have this patient population, you cannot assume that every patient is going to look the same, especially for CT pulmonary angiography. It's a huge difference when you actually deal with bolus timing versus just guessing when you should do scanning. When you talk about volume, you also have to talk about rate of injection, which is very important.

As you decrease the time of your scanning, do you have to increase the rate of injection? You can't keep the rate of injections the same because you cannot deliver the iodine, as Dr. Foley said. You're trying to attain a certain attenuation in the aorta, and to do that you need a good tight bolus to get that study. So you must increase your rate of injection.

SM: Is there any reason that we can't do that, or is there a safety issue?

VY: No, not really. I'm sure most people would agree that 4 mL/sec is done easily and it's not really an issue.

SM: I'd like to talk about 370 contrast material. Are there applications in your practice where you think you absolutely need to use that contrast concentration, or even higher?

WS: Yes, in cardiac studies, particularly in subjects with low ejection rates and timing problems. We've tried to do lesser concentrations, but I think that we need the higher concentration in the coronary arteries. The question is, if you increase that concentration, will you start to obliterate structures that you need to see? I think the jury is still out on that.

SM: That's a good point. Is there a cutoff point where we go so high that we start to lose information? You mentioned in your talk that there were some specific problems with getting the proper dosing. But for coronary scanning specifically, you're pretty convinced that 370 is the way to go?

WS: At least in our experience, I don't think that concentrations less than that are really going to be able to define the edges and things that we want to see that well.

JOHN HIBBELN, MD: I think that changing from a 300 to a 370 offers incremental advantages. It is not an absolute, it's relative with the relative higher amounts of iodine delivered. It's not a dichotomous situation.

SM: I wonder if you'll feel differently when you get your 8-channel scanner!

JH: Well, we'll see. But I think there is a definite incremental advantage, however. We recently performed a similar experiment as Rendon. We maintained all of the other parameters the same but switched the contrast from our current 370 back to our previous 300. Then, we retrospectively reviewed patients who had been receiving serial examinations for follow-up of metastases. There was generally a marked increase (with 370) in aortic density, in aortic opacification. Then there was a noticeable increase within the measured density within the livers and spleens on the scans with 370 versus those with 300. In so far as increasing contrast is good and increasing conspicuity is good, there is a definite incremental advantage.

SM: What do you use in liver imaging for metastasis?

JH: We use 370. We've been using 370 consistently for a number of years and we're very pleased with the results.

DF: My prediction will be that 370 will be of benefit in two areas. Number one, when you're looking for segmental vessel disease in kidneys or peripheral vessel disease in the liver, you're looking at smaller diameter vessels. You want to get a relatively high intra-arterial iodine to complement the isotropic voxel that you want to achieve with your 16-channel 0.625-mm detector system. In that circumstance, I think high intra-arterial iodine is going to pay a benefit and you're going to achieve that better with a higher concentration that you inject intravenously. I believe the other advantage of the 370 will be in the vascular phase--whether it's branch vessels in the trunk or it's branch vessels in the periphery of the body that you want to see.

SM: So as we move our CTA to the periphery, we need to keep our iodine concentration up there.

DF: Yes, I think you've got to look at what we are currently doing with CTA that we could do better, so we could replace even more catheter-based diagnostic arteriography.

RN: Well, I agree with Dennis. I would say that the major advantages are looking at small vessels. It may not be just for small vessels anatomically, but we may find that with appropriate high concentrations we can actually see lesions better because of small contiguous vessels. It's kind of like putting a catheter in the hepatic artery--clearly we know that additional lesions will visualize. If we get closer to that in a noninvasive method, then we're making some progress.

SM: Excellent analogy.

VY: I agree with Dr. Foley. I think peripheral CTA is the key application for 370. The more signal you want to get, the more useful 370 becomes. But it also depends on the application. For CT pulmonary angiography, I don't think 370 is going to help. It may actually create some artifacts from the venous side. But I agree that for peripheral CTA, it is going to be a key advantage.

PS: I'm going to take a very different approach. With regard to high concentrations, be it 370, 400, 425, or even higher, let me look into my crystal ball and tell you what I think. Why not use high concentration universally? I don't see any problem with it--why deliver volume if you don't have to? If I'm paying for contrast by volume, then categorically across the board with the exception of the artifacts, I would rather give a lower volume of higher concentration contrast and pass those savings on to my patients.

As scanners get faster and faster, I will want a higher concentration of contrast. I want to administer a bolus of 50 mL to get the exam done. It's a 4-slice detector today, an 8-slice detector pretty soon, so it's going to be 16 slices, 32, then 64. I don't want to be giving contrast when the patient's driving home. Give me 50 mL. I'll give that, then I'll get 100 images in one second and the patient's examination is done. There is no question in my mind that in 8 or 10 years CT scans are going to be a snapshot.

SM: If you're going to prevent having an option in your car for an IV injector, you're going to have to go that way.

PS: But I think viscosity is an issue and I think contrast companies who are looking to R&D efforts many years down the line ought to be aggressively looking at higher concentration, lower viscosity agents.

SM: The best of all possible worlds then is that scenario. Sergio, you gave a very nice presentation about safety. I just want to make sure that I understand that you found no difference in safety between 300 and 370 in nephrotoxicity in both at-risk patients and not at-risk patients in terms of both previous history and underlying conditions for nephrotoxicity. I know you weren't looking at contrast site injection issues. But you found no difference in safety issues in a huge population?

SERGIO GINALDI, MD: In the 300 and 370 products, that's correct, no difference. If I may, I want to echo what has just been said. We want to keep the volume low because we want to be competitive with angiographers. That's a clear issue to those of us that are not in academics. If we could have a study that is diagnostic and employs the use of a smaller amount of contrast, we certainly would become more competitive with interventional cardiologists.

As far as 370 versus 300, in our outpatient facility we use both, and for the study, we used it randomly. We don't preselect patients for one or the other. We may use one concentration one month, one group of products versus another group.

LT : In terms of the original 370 question, we use it essentially in a universal fashion across the board, except for some circumstances in which we just do not need a lot of iodine and speed's not important. As I mentioned in my presentation, it offers incremental benefits across the board. It's not a sudden all-or-none, but it does give incremental benefits.

There are artifacts in certain circumstances and there are ways to get around those. We scan bottom-to-top on our pulmonary angiograms to deal with that and there are preproduction algorithms that come from the manufacturers. The benefits even in pulmonary scanning are important when you start to look further out into the branches. The smaller a structure is, the higher the contrast resolution needs to be, the denser it needs to be for you to see it. There are some unique areas in which we care about this, specifically CT angiography and perfusion CT, whether that's in the body or in the head. The greater the difference between baseline and peak, the greater resolution we'll see on our studies. So, 370 does offer incremental benefits across the board, some unique benefits. But overall, it is a solid approach to routine imaging.

SM: That is an excellent summary of a lot of the others' opinions. Dr. Nelson, you said something in your presentation that didn't quite register with me and maybe you can clarify it. You analyzed the patients who received Isovue 370 and said you used 37 grams of iodine in that dosing. Then with the Isovue 300 you used 45 grams and didn't see a whole lot of difference, except in fall-off of aortic density. It's not quite the same amount of iodine, is it? So what do you think would have happened if you'd actually made it equal amounts?

RN: We chose 100 mL of 370 rather than 125 simply because that represents a cost savings. Fundamentally, I don't have any desire to go down. As a liver imager, I want the brightest liver because I know that even with the best CT and MR techniques, we're still missing 15% or 20% of lesions in noncirrhotic livers. In cirrhotic livers, we're probably missing 35% to 40% of the lesions; and that doesn't sit very well with me. So I still like high concentrations and high volumes. If we were to use 125 mL, we would have been closer to what the other protocol was, which was 150 mL of 300. But by doing those two comparisons you don't really gain anything in cost savings.

SM: The point about aortic density falling off, is that something we need to worry about?

RN: That's one reason why we included the qualitative arm in that study. We wanted to see whether or not the individuals who read those studies thought there was a diagnostic perception at least, and they did not. So I don't think that aortic drop-off in the venous phase is a significant factor from a diagnostic standpoint.

SM: Today, we have reviewed a lot of interesting experiences and some controversies related to the use of high-concentration contrast material. It's been extremely educational for me, since we've recently acquired multidetector CT scanning. Of course, we will all be getting into that as we upgrade our systems and we will want and need to know this information.

A lot of you will be conducting research studies that will help define some of the final answers as to which direction we're going to be moving in. We need to have industry work very carefully with us because while it seems as though we're moving in the right direction, there are certainly more improvements that should be explored, such as even higher-concentration iodine agents, with lower viscosity, although I'm not sure the physics are achievable. It's certainly a goal worth working toward.

At the moment, we have a grand opportunity here. We have the potential to get better imaging quality from our studies in less time, with the same level of patient safety. It is rare that a situation meets all those objectives simultaneously.

I want to thank each of you for your excellent presentations and your participation. We've given our readers an opportunity to really engage in this topic and learn a lot from your comments. Thank you very much.