This publication was supported by an educational grant from Amersham Health, Princeton, NJ. The opinions expressed in this publication are those of the authors and not necessarily those of Amersham Health.

Dr. Bae reports relationships with Tyco Healthcare and Mallinckrodt through patent agreements and as a consultant. Dr. Fishman reports relationships with Siemens Medical Solutions and Amersham Health as a consultant. Dr. Foley reports a relationship with GE Medical Systems through an investigator agreement. Dr. Naidich reports a relationship with Siemens Medical Solutions through its Advisory Board and as a consultant. Dr. Saini reports a relationship with GE Medical Systems through research support. Dr. Becker, Dr. Sahani, Dr. Siegel, Dr. Tahktani, and Dr. Zinreich report that no such relationships exist.

Dr. Foley directs the Section of Digital Imaging at the Medical College of Wisconsin, Milwaukee, WI.

As multidetector CT technology has evolved, so has our approach to contrast administration. This article will discuss the relationship between technology and contrast delivery, and explore its implications for pre-intervention planning.

The evolution in CT technology from 4- to 8- to 16-detector channels has led to an increase in beam width (Figure 1). When going from a 4-channel scanner (GE Light Speed Plus, GE Medical Systems, Waukesha, WI) to an 8-channel scanner (GE Light Speed Ultra), increasing the beam width from 5 to 10 mm doubles the speed of scanning, assuming detector collimation, rotation speed, and pitch remain constant. When going from an 8-channel (GE Light Speed Ultra) to a 16-channel scanner (GE Electric Light Speed 16), retention of the same 10-mm beam width, coupled with a change in detector configuration from 8 * 1.25 mm to 16 * 0.625 mm, results in a doubling of z-axis resolution for the same scan speed. (Compared with a 4-channel scanner, the same 16-channel configuration doubles both z-axis resolution and scan speed.) Alternatively, maintaining detector collimation constant at 1.25 mm increases the beam width of the 16-channel scanner to 20 mm. In this case, scan speed is twice as fast with a16-channel scanner as with an 8-channel scanner, and 4 times as fast as with a 4-channel scanner.

The general principles governing our approach to multidetector CT angiography (CTA) are as follows:

* We aim for an arterial enhancement of 250 to 300 HU. One important consideration is the need to distinguish arterial enhancement from calcification in the vessel wall, which is achieved by increasing the window width to 750 and the window level to 150 on the axial images, with appropriate presets on the maximum-intensity projection displays.

* In determining cephalocaudad coverage, we take into account both the need to acquire images during the first
circulation of contrast and the patient's ability to sustain a breath-hold. We generally consider 20 seconds a reasonable breath-hold interval.

* To time image acquisition, we inject a mini-bolus of contrast and observe aortic arrival time. We do not use bolus-tracking software, as our technology does not allow a sufficiently rapid transition from the detection of threshold contrast enhancement to the initiation of scanning.

Aortoiliac CTA

Aortoiliac CTA covers the territory from just above the celiac axis to the proximal thigh, approximately 30 cm. Figure 2 illustrates the timing of contrast administration and image acquisition using a 4-channel scanner. With a detector collimation of 1.25 mm, a pitch of 1.5, and a table speed of 15 mm/sec, the image acquisition interval is 20 seconds. Contrast material is injected at
5 mL/sec for 20 seconds, for a total of 100 mL (orange bar). In this schematic, image acquisition begins at 15 seconds (yellow bar), as determined by a preliminary mini-bolus, and the acquisition interval is equal in length to the injection interval. The injection-to-scan delay will vary from 12 to 30 seconds, depending on the individual patient's circulation time.

With an 8-channel scanner, z-axis resolution and scan rotation speed remain the same as with a 4-channel scanner (Figure 3). Beam width is doubled, however. If at the same time the pitch is reduced to 1.35, the acquisition interval is reduced to 12 seconds for the same cephalocaudad coverage. Contrast material is injected at 5 mL/sec for 12 seconds, for a total of 60 mL. Once again, the linkage between injection and acquisition is determined by the mini-bolus technique.

Figure 4 shows curved planar reformations of the same patient studied with a 4-channel and an 8-channel CT scanner on separate occasions using the same z-axis resolution. Attenuation of the aorta and iliac system throughout the cephalocaudad coverage is equivalent, despite contrast volume having been reduced by nearly half.

With a 16-channel scanner, the best approach to aortoiliac studies is to maintain the same beam width as with an 8-channel scanner and decrease detector collimation to 0.625 mm (Figure 5). If the pitch remains 1.35, the acquisition interval remains 12 seconds, and cephalocaudad coverage is unchanged. The z-axis resolution is doubled, however. Contrast is injected at 5 mL/sec for a total of 60 mL.

Figure 6 shows a patient with an aortoiliac stent graft. Images were acquired on an 8-channel scanner using a slice width of 1.25 mm, and a 16-channel scanner using a slice width of 0.625 mm. The effect of improved z-axis resolution is clearly evident in the superior edge definition of the right renal artery.

Stent-graft planning

Important issues in planning for stent-graft surgery include the ease of iliac access, the transverse dimensions of the superior neck of the aneurysm, the length of the superior neck (the proximal stent-graft placement zone), the angle of the neck in relation to the aneurysm, and the diameter and length of the distal landing zone in the common iliac arteries.

CT angiography can noninvasively obtain both the length measurements, which would otherwise be obtained by digital subtraction angiography with a calibrated catheter, and the diameter measurements, which would otherwise be obtained by intravascular ultrasound. Using center-line tracking, seed points, and automated edge detection, CT models the aorta and iliac arteries in three dimensions. It provides estimates of the true vessel length and the diameter, which is measured perpendicular to the vessel rather than to the body.

CT also plays a key role in postsurgical follow-up and the evaluation of stent-graft complications. Potential complications include endoleak (Figure 7), which might result in aneurysm expansion; graft thrombosis; stent graft migration or kinking; and branch vessel occlusion.

Renal studies

CT angiography is widely accepted for the evaluation of potential renal donors. With a 0.625-mm detector collimation, the 16-channel scanner provides detailed information on accessory vessels, points of branching, and renal artery stenosis (Figure 8). With a second-pass image acquisition, it can also detect renal vein anomalies.

In patients with marginal renal function, the 16-channel can be used as a super 8-scanner by increasing detector collimation to 1.25 mm and thereby doubling scan speed. Under such circumstances, the acquisition interval is reduced to 6 seconds, and contrast material is injected at 5 to 6 mL/sec for 6 seconds, for a total contrast volume of just 30 to 36 mL. One might wonder whether it is possible to accurately time image acquisition to coincide with the passage of the contrast bolus, but with the aid of the preliminary mini-bolus, we have been successful in doing so.

Figure 9 shows an extension of an aortoiliac study into the upper thigh of a patient with bilateral renal artery implants, bilateral internal ureteral stents, a femoral-femoral artery crossover graft, and serum creatinine level of 1.5 to 2 mg/dL. This study was done with 40 mL of contrast material.

Table 1 summarizes the comparison of 4-, 8-, and 16-channel scanners for aortoiliac CTA.

Thoracoabdominal aortic CTA

There are two approaches to imaging the thoracoabdominal aorta with the 16-channel scanner. Detector collimation can be set at 1.25 mm for a full 20-mm beam width. In this case, the acquisition interval for the thoracoabdominal aorta is equivalent to that for an aortoiliac study, despite a cephalocaudad coverage of 60 cm. Contrast volume is approximately 60 mL. The alternative approach is to use a detector collimation of 0.625 mm. In this case, both the acquisition interval and the contrast load will double.

The choice between the two ap-proaches is guided by the patient's cardiovascular status and renal function, as well as the indication for the study. For pre-operative planning in a patient with a suspected thoracoabdominal aneurysm, it is essential to obtain detailed information on small vessels that parallel the slice plane. Therefore, it would be appropriate to select a detector collimation of 0.625 mm and a contrast volume of 120 mL. In the case of aortic dissection, where small-vessel detail is less important, a detector collimation of 1.25 mm is sufficient, and contrast volume can be reduced to 60 mL.

Extremities

Patients are usually referred for CTA of the extremities in preparation for vascular or plastic surgery, or for evaluation of orthopedic problems. Multidetector CT is both suited for and challenged by the demands of imaging the extremities.

The distance from the renal vascular pedicle to the feet is approximately 120 cm. To cover this distance with a 4-channel scanner, it is necessary to use a 2.5-mm detector collimation, a pitch of 1.5, and a table speed of 3 cm/sec, resulting in a 40-second acquisition interval. Contrast material is injected at 4 mL/sec for 40 seconds, and a mini-bolus of contrast material determines the timing of image acquisition.

Lower-extremity angiography becomes more complex in patients with very slow runoff, particularly below the knee. Under such circumstances, we delay initiation of image acquisition beyond the aortic arrival time as determined from the contrast mini-bolus, in order to avoid scanning faster than the contrast circulation.

With an 8-channel scanner, we reduce detector collimation to 1.25 mm. With no changes in the other scan settings, the image acquisition interval remains about the same, 44 seconds. With a 16-channel scanner, we generally maintain detector collimation at 1.25 mm, which reduces the acquisition interval from 44 to 22 seconds. We inject contrast material at 5 mL/sec, for a total contrast volume of 110 to 120 mL. This volume of contrast material is acceptable in patients with marginal renal function or substantial cardiovascular disease.

Table 2 summarizes differences in the scan protocols for CTA of the extremities using 4-, 8-, and 16-channel scanners.

A key issue in lower-extremity CTA is the need to determine the degree of stenosis in areas with significant calcified plaque. This problem necessitates the use of curved planar reformations and automated measurement techniques, to determine the diameter and area at the sites of stenosis (Figure 10).

Image processing remains a key issue, particularly in extremity CTA. It is important, for example, to be able to do bone segmentation at the workstation effectively and quickly. Vessel tracking, particularly below the knee, and automated stenosis sizing are also critical in making CTA of the extremities a robust and acceptable technique that will replace other forms of angiography.

Conclusion

Intravenous CTA of the thoracoabdominal aorta and abdominal visceral vessels is a robust technique that provides all the requisite diagnostic information for pretherapy planning and postintervention follow-up. Improved performances associated with the progression from 4- to 8- to 16-channel scanners is, in each instance, associated with either faster acquisition, improved resolution, or reduced contrast load. Intravenous CTA of the extremities can be used for preoperative evaluation of iliac, femoral, and popliteal arterial disease. Display of isolated tibio peroneal arterial disease can be more problematic.

Figure Captions

FIGURE 1. Matrix detector technology evolution. The beam width of an 8-channel CT scanner, at 10 mm, is twice as large as that of a 4-channel CT scanner. Detector collimation, at 1.25 mm, remains unchanged, resulting in a doubling of scan speed. With a 16-channel scanner, the beam width can remain 10 mm while detector collimation can be reduced to 0.625 mm, doubling z-axis resolution. Alternatively, beam width can be increased to 20 mm while detector collimation remains 1.25 mm, doubling scan speed in comparison to an 8-channel scanner, or quadrupling it in comparison to a 4-channel scanner.

FIGURE 2. Timing of contrast administration and image acquisition during aortoiliac CTA using a 4-channel scanner. Detector collimation is 1.25 mm, pitch (Z) is 1.5, and table speed 15 mm/sec (7.5 mm per scan rotation/0.5 sec per rotation). The image acquisition interval is 20 seconds. Contrast material is injected at 5 mL/sec for 20 sec, for a total of 100 mL (orange bar). Image acquisition begins 15 seconds after the beginning of contrast injection (yellow bar), as determined by a preliminary mini-bolus for this hypothetical patient. (Note: The injection-to-scan delay will vary from 12 to 30 seconds, depending on individual patient ' s circulation time.) The acquisition interval is equal to the injection interval.

FIGURE 3. Timing of contrast administration and image acquisition during aortoiliac CTA using an 8-channel scanner. Scan rotation speed and z-axis resolution are the same as with a 4-channel scanner. Beam width is doubled, however. If pitch (Z) is reduced to 1.35, the acquisition interval is reduced to 12 seconds. Contrast material is injected at 5 mL/sec for 12 sec, for a total contrast volume of 60 mL.

FIGURE 4 . Curved planar reformations from the same patient studied with both (A) a 4- and (B) an 8-channel CT scanner on separate occasions using the same z-axis resolution. Attenuation of the aorta and iliac system throughout the cephalocaudad coverage is equivalent, despite contrast volume having been reduced by nearly half.

FIGURE 5. Timing of contrast administration and image acquisition during aortoiliac CTA using a 16-channel scanner. Reducing detector collimation to 0.625 mm doubles z-axis resolution. Acquisition interval is 12 sec, and contrast material is injected at 5 mL/sec for a total of 60 mL.

FIGURE 6. Patient with an aortoiliac stent graft. Images were acquired on (A) an 8-channel scanner using a slice width of 1.25 mm, and (B) a 16-channel scanner using a slice width of 0.625 mm. The effect of improved z-axis resolution is clearly evident in the superior edge definition of the right renal artery in the image acquired on the 16-channel scanner.

FIGURE 7 . (A) Curved planar reformation of a patient with a Type 2 endoleak (arrow). (B) Due to graft limb thrombosis (large arrow), the patient has had a femoral-femoral bypass graft (small arrows).

FIGURE 8. Potential renal donor. With a 0.625-mm detector collimation, the 16-channel scanner can provide detailed information on accessory vessels, points of branching, and renal artery stenosis.

FIGURE 9. Extension of an aortoiliac study into the upper thigh of a patient with bilateral renal artery implants, bilateral internal ureteral stents, an axillofemoral and femoral-femoral artery crossover graft, and serum creatinine level of 1.5 to 2 mg/dL. This study was done with 40 mL of contrast material.

FIGURE 10. (A) Maximum-intensity projection shows heavy calcification in the distal aorta, iliac arteries, and the femoral arteries down to the knee. (B and C) Curved planar reformations show the internal lumens of the right and left iliac, femoral, and popliteal arteries without obscuration by the heavy arterial wall calcification. The left superficial femoral artery is occluded and distal runoff is via the left profunda femoris artery with collateral reconstitution of the left popliteal artery.

Discussion

ELLIOT K. FISHMAN, MD: Thank you, Dennis. Any questions?

KYONGTAE T. BAE, MD, PhD: You said you like to have desired arterial enhancement of 250 to 300 HU. But then with some of the reduced contrast volume, you are not going to achieve that, right?

FOLEY: That's a good question. The issue relates to the fact that we are not really dealing with recirculation and humping up the aortic profile. So if you have a very short acquisition of 6 seconds, you don't get the recirculation effect. I haven't gone systematically to look at those patients, but they probably don't have as much aortic attenuation as we have been talking about.

BAE: Just from looking at the image, it looks like enhancement is not at the level of 250 HU.

FOLEY: Yes, but the compensation for that is--instead of injecting at 5 mL/sec--inject at 6, and maybe inject at 7 mL/sec. One thing I didn't indicate on that schematic is that, actually, when we go to a 6-second acquisition we increase the in-jection rate to 6 mL/sec instead of 5.

SANJAY SAINI, MD: Also, when you have a longer coverage, why not go to a 1.75 pitch? You stayed at 1.35.

FOLEY: That's a good point. With regard to the vessel detail, if you're looking at a vessel that's perpendicular to the slice plane, it probably would not actually diminish anything at all, that would be an advantage.

FISHMAN: In terms of CTA, outside of the heart, the most challenging is definitely the lower extremity in terms of timing. You said that if you were using the automated bolus tracking, you measured the lower abdominal aorta. What values do you use for that?

FOLEY: Actually, we are still using the mini-bolus technique. So we don't use the thresholding. I'd be interested in other people's opinions.

FISHMAN: But even with the mini-bolus, there is the issue. Trauma patients with glove/ring injuries are easy to image. When you evaluate a 60-year-old for atherosclerotic disease, what's good for the left is not good for the right. What do you do in those situations?

FOLEY: We haven't had a sufficient experience to determine that.When you have iliac occlusive disease or femoral occlusive disease, generally the collateral circulation that you fill out with an intravenous injection overcomes the problem that you have if you do selective arterial injections with digital subtraction angiography. But you're right, below the thigh and below the knee, it's going to be an issue. You just hope that you have such a long bolus, and you've got some latitude between a long bolus and a long acquisition interval, that you're not going to run into a problem.

I've heard other people present material on CTA of the lower extremities who have not really indicated there's a major problem with different flow rates between the two sides.

FISHMAN: But, you did a lot of angiography, and you know from that there's a significant problem. We do sort of a crude thing: I just watch the images in all those CTAs. If I see that the left looks good and the right doesn't, I just wait 20 more seconds, and scan again. I think it's impossible for patients to be perfect on both.

CHRISTOPH R. BECKER, MD: What would be a good idea, I've heard, is to test bolus in the femoral artery, to see when the arrival is there, and to then concentrate on the diseased leg.

FOLEY: Yes, that's a good point.