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. Siegel is a Professor of Radiology and Pediatrics at the Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO.

Multidetector computed tomography (CT) offers the same advantages in children as it does in adults. These advantages include faster scan acquisition, increased resolution, and optimal three-dimensional (3D) processing. The implications of multidetector technology are often different in pediatric patients, however. ¹

Faster scanning improves contrast enhancement in any population. But in children, it also reduces the need for sedation. With single-slice scanners, sedation is needed in 40% to 100% of pediatric patients. Early experience with 16-row multidetector scanners suggests that although sedation cannot be completely eliminated, it is now needed in <5% of pediatric patients. In addition, respiratory artifacts have nearly been eliminated. ²

Improvements in resolution and 3D reconstructions offer additional advantages. Together they make CT an excellent alternative to magnetic resonance (MR) imaging. Although exposure to ionizing radiation puts CT at a relative disadvantage when compared with MR, newer techniques are reducing the radiation dose. More important, children under 5 years of age require sedation during hour-long MR examinations. The risks of sedation are real, and in some instances they may exceed the risks of exposure to ionizing radiation.

At our institution, CT angiography (CTA) is beginning to replace conventional catheter angiography for evaluation of cardiac lesions. ³ It is associated with a radiation dose at least 2 to 3 times lower than that of conventional angiography. This estimate is based on older radiographic parameters, including a kV of 140 and an mA of up to 400. With more contemporary radiographic settings
(80 kV and mA adjusted for weight), the radiation dose is likely to be much lower, perhaps 8 to 10 times lower than that of conventional angiography.

Achieving the best result

Optimizing CTA in children requires attention to scanner design. It is possible to perform CTA using a 4-row scanner, but a 16-row scanner is preferred. Patient preparation is also important. We place the intravenous (IV) line before the patient arrives in the imaging suite. Waiting to place the IV line until the child is on the CT table is likely to result in an uncooperative child.

We use the largest-gauge catheter possible, generally a 22-gauge, and insert it in the antecubital region if possible. Some patients arrive with an IV line already placed in a scalp vein or a catheter in the foot, however. For an abdominal scan, oral contrast must be administered before the study.

Table 1 outlines technical factors to consider in CTA of pediatric patients. The selection of slice width usually represents a compromise. With our 4-row Siemens scanner (Siemens Medical Solutions, Iselin, NJ), detector collimation thickness can range from 1 to 5 mm, or greater. A slice width of 1 mm will produce images of superb resolution but at an excessively high radiation dose. A slice width of
5 mm will reduce the radiation dose, but resolution may be unacceptable. A slice width of 2.5 mm represents a reasonable compromise. Table speed is as fast as possible, 15 to 20 mm/sec, also with the goal of reducing radiation exposure. Scan speed is 0.5 seconds.

With our Siemens 16-row scanner, slice width can be 0.75 or 1.5 mm. We use the larger width to reduce radiation exposure. Table speed is 24 to 36 mm/sec, and scan speed is 0.5 seconds. With both 4- and 16-row scanners, the slice reconstruction thickness is 3 to 5 mm for routine viewing.

It is important to tailor the selection of radiographic parameters to patient size, as shown in Table 2. In the case of a neonate, whose weight is <15 kg, the maximum allowable mA in the chest is approximately 25, and in the abdomen, approximately 30. These limits gradually increase until the patient reaches approximately 50 kg, when it becomes possible to use adult radiographic standards. Another factor to consider is kVp. In a child of <50 kg, we successfully use a kVp of 80. In a child who weighs >50 kg, we increase the kVp to 100 to 120.

The multidetector CT study itself is very short. We do not routinely perform multiphasic imaging. In the chest, we usually acquire images in the arterial phase, and in the abdomen, the portal venous phase. Our approach varies with the clinical indication; however, this topic is beyond the scope of this article.

Multiphasic imaging is justified under certain circumstances. In complicated cases of congenital heart disease, it may be helpful to do both arterial- and venous-phase imaging. In evaluating hepatic tumors, we perform dual-phase imaging as well.

Contrast administration

The key issues of contrast administration are similar in children and adults: contrast volume, flow rate, and the timing of image acquisition. We use nonionic contrast media and a standard concentration of 300 mgI/mL. In the past, the typical contrast dose in children was 2 mL/kg, but faster scanners permit a reduction to 1.5 mL/kg. The maximum contrast dose in children is the lesser of 4 mL/kg or 125 mL. In our experience with faster scanners, it is very rare to need more than 2 mL/kg.

If the patient has a peripheral IV or a central line, we inject contrast by hand, delivering the dose as quickly as we can. In such cases, the speed of injection will vary by the location of the access line and the caliber of the needle. ⁴ If it has been possible to insert a 22-gauge or larger catheter in the antecubital space, we use a power injector to deliver contrast material at a flow rate of 1.5 to 2 mL/sec. With a 20-gauge catheter, the injection rate can be increased to 2 to 3 mL/sec. With a 24-gauge catheter or a central line, contrast is injected at 1 mL/sec.

Timing

The timing of image acquisition is the final and critical step in optimizing contrast delivery. One of the biggest problems with smaller-caliber needles is that the scan may be completed before all of the contrast has been injected. Therefore, when imaging vascular structures in the chest in children weighing <10 kg, we use a scan delay of 12 to 15 seconds.

In larger children, we sometimes use a scan delay of 20 to 25 seconds for imaging vascular structures in the chest. We prefer to use automated scan initiation or bolus tracking, however. In smaller children, this approach is not sufficiently reliable. With a contrast volume of only 4 to 8 mL, it can be difficult to achieve the enhancement threshold necessary to trigger the scan, 100 HU in chest studies.

We sometimes use electrocardiographic gating in adolescent patients, particularly if we suspect a very small lesion, such as a septal defect, or some type of a postoperative small conduit, such as the baffle or Fontan grafts. We seldom use ECG gating in neonates because the heart rate is so fast, and gating increases radiation dose.

For CTA of the abdomen in neonates weighing <10 kg, the scan delay typically is 10 to 15 seconds for image acquisition during the arterial phase, and approximately 45 seconds for image acquisition during the venous phase. In infants and children weighing >10 kg, scan delays of 20 to 25 seconds for the arterial phase, and 50 to 55 seconds for the venous phase are typical. It is also possible to use bolus tracking and trigger scanning at approximately 50 HU, but usually an empiric delay works well in abdominal studies.

Postprocessing

We use a two-dimensional (2D) multiplanar technique because it is easy and can be done at the console, where clinicians can also look at images quickly. This technique provides substantial information about the vasculature, but it lacks depth. Therefore, we prefer to use 3D volume rendering for postprocessing of data from CTA studies.

Volume rendering enables the visualization of both vessels and airways. ⁵ Vascular lesions in children are often associated with airway abnormalities, and volume rendering affords the option to study both. We also find that some of the maximum intensity projections, particularly the thick slabs, are useful for looking at intraparenchymal vessels and for examining arteriovenous malformations.

In reviewing volume-rendered datasets, it is important to use a workstation with real-time, interactive capabilities. It remains essential to review the axial images, as they often provide diagnostic information that is complementary to that derived from 3D images. In some cases, a diagnosis can be made only on the basis of either a 2D or a 3D image.

Applications

Vascular

The vascular applications of pediatric CTA include the diagnosis of congenital abnormalities of the thoracic vessels, and vessel mapping for tumor staging and surgical planning. Examination for coarctation of the aorta, or a double aortic arch, are common. We also evaluate abnormalities of pulmonary venous return and extralobar sequestration.

Figure 1 depicts a 10-day-old infant girl with congestive heart failure. The most likely cause is coarctation of the aorta. CTA was performed using 8 mL of contrast and a 12-second scan delay. It is difficult to make a diagnosis on the axial view. On the multiplanar sagittal view and the 3D reconstruction, a very tight high-grade stenosis is evident, indicating the need for surgical repair.

Figure 2 illustrates the use of CTA to differentiate extra-lobar sequestration from neuroblastoma. In this newborn,
8 mL of contrast was slowly injected through an existing IV line in the scalp. The axial images, though not optimal, demonstrate a tiny vessel arising from the aorta. A tiny vein draining the sequestration was recognizable only in retrospect. On multiplanar 3D images, the arterial feeder that characterizes sequestration is easily visualized, however. An unusual vein providing drainage to the portal vein is also evident. This case demonstrates that even with a very small amount of contrast injected through a peripheral IV line, CTA produces clinically useful images.

Examination of sacrococcygeal teratoma in a neonate represents another vascular application of CTA, in this case for tumor staging. Often the axial view will easily show the tumor, but the vessels may be barely visible. By comparison, on the 3D reconstruction, a feeder vessel may be evident coming off the middle sacrococcygeal artery.

Cardiac

One of the most exciting uses of CTA is in cardiac applications, including the diagnosis of congenital shunt lesions and the evaluation of postoperative anatomy, including palliative shunts and complex heart disease.

Figure 3 presents the example of a 2-week-old infant who was cyanotic and thought to have congenital heart disease. CTA, accomplished with 8 mL of contrast material and mAs of 30, clearly demonstrates a right hemitruncus, in which one pulmonary artery arises from the aorta and the other from a pulmonary trunk.

Another cardiac application of CTA, the postoperative assessment of a patient with a history of teratology of Fallot, is depicted in Figure 4. A surgically placed shunt between the subclavian artery and the pulmonary artery has thrombosed, as shown on both axial images and multiplanar reconstruction.

Finally, Figure 5 shows the postoperative evaluation of a patient who, as a neonate, underwent repair of transposition of the great vessels. CTA easily reveals a 5-mm graft between the right atrium (RA) and left ventricle, as well as a small septal defect between the right and left ventricles.

Safety

CTA has many exciting applications in pediatric patients, and many advantages. The use of 16-slice CTA offers the opportunity to eliminate the long periods of sedation associated with MR, and reduces the radiation exposure associated with conventional angiography.

The benefits of accurate diagnosis alone generally outweigh the risks associated with CTA. Nonetheless, care must be taken to reduce unnecessary risks, as children <10 years of age are more sensitive to the effects of radiation than are middle-aged adults. To reduce radiation exposure, the CTA protocol and technical settings must be optimized. The tube current and kVp must be as low as possible, the table feed as fast as possible, the number of contrast-enhanced phases must be minimized, and automated dose reduction technology should be used, if it is available. Inappropriate referrals to CT must also be eliminated. ^6,7

Conclusion

As CT technology advances, it is having an increasingly profound impact on the care and imaging of children. When performing CTA in children, study design and execution are critical. Even with small volumes of contrast and very small anatomy, CTA is successful if performed with an attention to detail. Patient preparation is crucial, as are the proper selection of technical factors and optimal delivery of contrast material. Data processing, particularly the creation of 3D volume renderings, is often essential in making a diagnosis.

Figure Captions

FIGURE 1. Aortic coarctation in a 10-day-old girl with congestive heart failure (8 mL contrast, 12-second delay time). (A) The diagnosis of coarctation of the aorta is difficult on the axial view, but on (B) the multiplanar sagittal view and (C) the 3D reconstruction, a very tight high-grade stenosis is evident (arrow in B), indicating the need for surgical repair.

FIGURE 2. Differentiation of extralobar sequestration from neuroblastoma in a newborn (slow scalp vein injection, 8 mL contrast). (A and B) Axial images demonstrate a tiny feeding artery (arrow in A). A tiny draining vein (arrow in B) was recognizable only in retrospect. (C, D, and E) On multiplanar 3D images, the arterial feeder that characterizes sequestration is visualized easily. An unusual vein providing drainage to the portal vein is also evident.

FIGURE 3. (A and B) CTA provides the diagnosis for a right hemitruncus in a 2-week-old infant (30 mAs, 8 mL contrast). One pulmonary artery [PA] arises from the aorta [A] and the other from a pulmonary trunk.

FIGURE 4. Postoperative assessment of a palliative shunt in an adolescent patient using bolus-tracking technique. In this patient with a history of tetralogy of Fallot, a surgically placed shunt (arrows) between the subclavian artery and the pulmonary artery has thrombosed, as shown on both (A and B) axial images and (C) multiplanar reconstruction.

FIGURE 5. Postoperative assessment of a patient with repaired transposition of the great vessels. (A) CTA easily reveals a 5-mm graft between the right atrium [RA] and the left ventricle (arrow), as well as (B) the shunting of contrast from the right to the left ventricle through a tiny residual ventricular septal defect (arrow).

Discussion

ELLIOT K. FISHMAN, MD: Any questions?

KYONGTAE T. BAE, MD, PhD: You use pretty much the same delay for CTA of the chest and abdomen?

MARILYN SIEGEL, MD: Yes. Chest CTA is initiated at 12 to 15 seconds in very small patients and at 20 to 25 seconds with an automated program in larger patients. The major role of CTA in the abdomen is in the evaluation of the liver. The timing of the arterial phase is at 12 to 15 seconds in small patients and 20 to 25 seconds in larger patients. It works.

BAE: So 16-slice CT hasn't really changed the timing.

SIEGEL: No, the time delay for CTA of the chest and abdomen remain similar.

FISHMAN: Pediatric patients are obviously unique, in terms of selection of contrast concentration or types of contrast. Are there specific rules you recommend?

SIEGEL: We still use concentrations of 280 to 320 mgI/mL. This results in excellent vascular opacification as long as the catheter is of adequate size to sustain a fast flow rate.

DAVID P. NAIDICH, MD: Are kidneys in neonates more vulnerable?

SIEGEL: Yes. Stasis nepropathy is a problem in the neonate. This is the result of the accumulation of glycoproteins (the principal one being the Tamm-Horsfall protein) within the tubules, leading to temporary oliguria. The use of intravenous contrast material has been associated with the precipitation of excess amounts of Tamm-Horsfall protein. Stasis nephropathy is transient and generally resolves within the first week.

FISHMAN: Do you think using a higher concentration gives you more of a window doing a successful CTA? If you are looking at vessels or the hepatic artery or renal arteries, would a higher concentration give you more leeway in terms of finding a higher sweet spot or a wider sweet spot, or would it make a difference?

SIEGEL: If one were injecting through a very small needle, then a higher concentration of iodine might be beneficial. However, experience using higher concentrations of contrast material in children has not yet been reported in large series. As noted above, excellent contrast enhancement can generally be achieved as long as the catheter caliber is large enough to allow a fast flow rate. It also helps to have the catheter in an upper extremity, particularly in an antecubital location.

W. DENNIS FOLEY, MD: I think the answer to that question is that if your iodine load is the same, and your injection rates are the same, and one contrast agent is more concentrated than the other, then your imaging window is just going to vary in time. It's not going to have any wider temporal window. In other words, with a higher concentration under those circumstances, the number of iodine atoms going into the circulation is going in faster per second.

BAE: Can you check creatinine in this population, then administer hydration?

SIEGEL: Yes, but the volume of fluid that can be given to a neonate is limited.

NAIDICH: We're using iopromide, which is 370 mgI/mL, as a routine for pulmonary artery CT. It seems empirically that going from 300 to 370 mgI/mL has reduced the number of indeterminate CT studies that we've had. I appreciate that it's a large iodine load; but just the opposite of where you're leading with your question, I think we're still relatively safe with using a higher, rather than a lower, dose of iodine. It would be nice to reduce it, but I don't see, at this point, that we could do that.

FISHMAN: All right, thanks very much, Marilyn.