Applied Radiology

Lawrence N. Tanenbaum, MD
Section Chief, Neuroradiology, MRI & CT, New Jersey Neuroscience Institute, JFK Medical Center­Edison Imaging Associates; Assistant Professor, Neuroscience, Seton Hall University; Edison, NJ

Since the introduction of the first slip-ring scanners, neurological CT has been characterized by substantially shorter scan times, a trend that has accelerated with rapid advancements in multidetector systems. Today, we can complete a CT angiogram that extends from the thoracic aorta to the toes in less than one minute, and a CT of the neck in mere seconds.

Although the ability to scan more quickly has been the dominant advantage of recent improvements in CT technology, an increasing number of detector channels also presents the option to balance speed with improvements in image quality. Either approach can have a substantial impact on the selection of contrast agents, the volume used, and the administration method. Iodine concentration is of particular importance. In many applications of neurological CT, higher-concentration, lower-volume contrast media produces the best result. ^1,2

Scanning Parameters

With single-channel CT scanners, we conduct most neurological examinations at a pitch of 1.5. (Pitch is defined as the distance the table travels per tube revolution, divided by the detector cluster dimension.)

Under some circumstances, we select a higher pitch when using a single-detector scanner, for example when we need very rapid coverage in order to chase a small bolus of contrast, to image a long area of interest, or to scan an uncooperative patient. In general, however, a pitch of 1.5 remains our internal reference for image quality and slice profile.

With 4-channel systems, our ap-proach for most applications is to use a pitch of 3, which enables us to double scanning speed while improving the slice profile. With a 15-mm table incrementation per tube revolution and a collimation of 4 * 1.25 mm, it is possible to cover 150 mm in 10 revolutions in as little as 5 seconds. Under certain circumstances, such as CT angiography (CTA) from the aortic arch to the Circle of Willis and studies of the soft-tissue neck, we take advantage of the increased scan speed that a pitch of 6 provides. With a table speed of 30 mm per tube revolution, it is possible to cover 300 mm in 10 revolutions in the same 5 seconds, with an acceptable increase in slice profile.

An 8-channel system offers even more flexibility. Because of its greater speed, we can use a moderate pitch of 7 to 10.8 for most applications, including imaging the soft-tissue neck. Alternatively, we can push scan speed to the maximum the system can achieve in order to do CT angiography or studies that necessitate long breathholds, such as 1.25-mm collimation scans of the abdomen and pelvis. Alternatively, we reduce pitch to 5 for head CTA. Here speed is less critical and, instead, the primary objective is the acquisition of superior images of low contrast resolution and least possible image noise.

In addition to multiple detectors, CT scanners now also offer extreme-ly fast scan rotation speeds, ranging from 0.5 to 1 second. A scan rotation speed of 0.5 seconds, which we typically use for vascular opacification or CTA, yields twice the coverage for a given breathhold. This gives us an opportunity to be more effective in our utilization of contrast.

Optimizing Contrast

In neurological CT, one of our primary goals in using contrast is to achieve optimal depiction of the breakdown of the blood-brain barrier (Table 1). Detection of neurological disease is in part governed by the fact that, while normal brain tissue will not enhance, lesions that do not maintain the blood-brain barrier will. In other areas, such as the soft-tissue neck, we are looking for mucosal enhancement. Under other circumstances, such as CTA and neck CT, we want to maintain maximum vascular opacification. We also want to maximize image contrast resolution, particularly in perfusion imaging. ¹

One way to take advantage of today's extraordinary imaging speeds while still delivering the appropriate amount of iodine for a specific imaging application is to use higher injection rates. We use injection rates of 2 mL/sec for our routine applications, 3.5 mL/sec for CT angiography, and 4 mL/sec for perfusion CT.

In a busy outpatient setting, injection rates are limited by a number of practical concerns. These include extravasation risk, as well as the time it takes to obtain intravenous access suitable for high injection rates and its impact on throughput. We scan 4 patients every hour on our 8-channel system, and establishing a large-bore intravenous line can cause unacceptable delays.

Patient tolerance to contrast injection is also key in enabling the use of high injection rates. The local effects of contrast, such as a feeling of warmth or a burning sensation, must be considered, as must the potential for nausea and vomiting. It is important to remember that an interrupted exam is a lost exam. We feel most comfortable using nonionic contrast media for applications that require high injection rates, a choice that has been made progressively easier by the reduced cost differential between ionic and nonionic contrast agents.

As in most practices, we work with a variety of helical CT scanners, including 1-, 4-, and 8-channel systems. We hope to add a 16-channel system in the near future. To accommodate such a wide range of technology, our approach has been to match the concentration and volume of contrast to both the specific application and each scanner's capability.

In the soft-tissue neck, we achieve excellent opacification with a total iodine dose of about 23 grams, delivered as 75 mL of Isovue 300 (Bracco Diagnostics, Princeton, NJ) (Table 2). In the orbit, we use the same dose. In the head, we use 37 grams of iodine, delivered as 100 mL of Isovue 370. In perfusion imaging, the goal is to maximize the difference in attenuation between baseline and peak enhancement. It is in this type of application that a high-density contrast agent can have an enormous impact. We deliver approximately 15 grams of iodine in 40 mL of Isovue 370.

CT angiography is accomplished using a range of iodine doses. We do brain CTA with 50 mL of Isovue 370, achieving an excellent exam with high image quality using only 18 grams of iodine. When we do a full neurological angiogram, covering from the aortic arch through the intracranial vasculature, we use approximately 75 mL of Isovue 370 and deliver a total of 28 grams of iodine.

Neck

Neck CT presents an unusual set of challenges. The speed offered by multidetector helical CT is not necessarily helpful in this instance, as the mucosa and mucosal lesions need time to enhance before images are acquired. It is also desirable to limit the overall contrast dose to enhance patient tolerance and safety, and cost-effectiveness. In addition to imaging the mucosa, we scan the neck while vessels are still opacified to enable differentiation of lymph nodes from vascular structures.

We balance this competing set of goals with 23 grams of iodine administered in 75 mL of Isovue 300. We then tailor scanning parameters to suit each generation of CT scanner technology. With a single-channel scanner we use a pitch of 1.5, whereas with a 4-channel scanner we use a pitch of 6, and with an 8-channel scanner we use a pitch of 11. We image after a scan delay of 20 to 30 seconds.

We can take advantage of advances in technology by selecting slice thickness as well. On a 1-channel scanner, we use 3-mm collimation, whereas on a 4-channel system we use 2.5-mm collimation. On an 8-channel system it is possible to accomplish all of the competing goals of neck CT with 1.25-mm source images. The tighter the collimation the better and more seamless our multi-planar imaging.

Figure 1 offers a classic example of how neck CT can be optimized by balancing the capabilities of fast scanners and higher-concentration contrast media. The patient has squamous cell carcinoma at the tongue base and tonsillar pillar. A large lymph node is seen easily, even with poor vascular opacification. Note that smaller nodes are easily differentiated from small vascular branches because of the excellent contrast resolution between the vessels and the lymph nodes.

The Orbit

In imaging the orbit, contrast serves primarily to characterize lesions, and a moderate dose of a moderate-concentration contrast agent will suffice. An iodine dose of 23 grams, delivered as 75 mL of Isovue 300, accomplishes the task very well. Speed is not a key factor in obtaining high-quality images of the orbit. Accordingly, we use a 30-second delay between injection and imaging to allow time for enhancement. We use 1- to 1.25-mm collimation and then image at a pitch that is appropriate for the particular scanner: 1.5 with a single-channel system, 3 with a 4-channel system, and 7 with an 8-channel system. Because we are very conscious of radiation dose when we scan the orbit, on our multichannel scanners we obtain all of our studies with a single scan in the axial plane and reformat the coronal and, occasionally, the sagittal oblique planes.

Figure 2 is an example of an orbital pseudotumor. The enlarged medial rectus muscle on the right is contrasted with the normal appearing medial rectus on the left.

The Head

Head CT poses an interesting set of contrast-related challenges. Contrast dynamics are not critical, nor is speed. However, a dose-sensitivity relationship has been well-demonstrated in head CT. ^3,4

We use 100 mL of Isovue 370 to approximate a standard contrast dose, delivered as an intravenous drip. This approach is helpful on 2 fronts: It creates a delay between administration and imaging while the contrast is dripping in, which is critical to depiction of the abnormal blood-brain barrier, and it enables us to avoid the costs of using disposable supplies for injection.

Figure 3 provides an example of how lesion detection often hinges on imaging the abnormal blood-brain barrier with contrast. In this case, noncontrast images suggest an abnormality in the frontal lobe and, perhaps, the right cerebellar hemisphere. When the blood-brain barrier is imaged with the appropriate dose of contrast, a lesion in the right cerebellar hemisphere is enhanced, as well as one in the frontal lobe, differentiating tumor from edema. Figure 4 demonstrates a moderate-grade temporal lobe glioma.

Whether modern scanners improve lesion visualization has not been studied. It is reasonable to think they might, given their vastly superior contrast resolution, ability to acquire thinner slices, and reduction in beam hardening or partial volume artifacts in the base of the brain, middle cranial fossa, and posterior cranial fossa.

Figure 5 exemplifies how more advanced CT technology can produce better images. It is easy, for example, to see the tiny nodular metastasis in the right cerebellar hemisphere, as well as the enhancing metastasis on the left. With previous generations of CT scanners, it might not have been possible to detect any abnormality in the posterior fossa, as a result of artifact.

Perfusion Imaging

First-pass perfusion imaging of the brain is a fairly new application of contrast-enhanced CT. The concept is to monitor the first-pass bolus of iodinated contrast through the cerebral vasculature. The contrast bolus causes a transient rise in attenuation that is proportional to the amount of agent in a given region. Then, integration of data over the course of the first pass of the contrast agent enables the creation of pixel-by-pixel maps of brain perfusion.

Perfusion imaging is simple to do in the clinical setting and typically has its greatest role in acute stroke, although tumor is another potential application. The procedure involves first selecting a slice or slices covering 3 vascular territories. A limited amount of contrast, 40 mL, is injected at 4 mL/sec. We use software that compensates for the modest 4 mL/sec injection rate; otherwise it would be necessary to inject at 10 mL/sec to accomplish, in essence, an instantaneous arrival of the contrast agent at the brain.

Since the goal of contrast administration is to maximize the difference between baseline and peak enhancement, the highest-concentration contrast agent should be the most effective, at least in theory. We use Isovue 370 and the scan takes 45 seconds.

The clinical utility of perfusion imaging can be compelling. Figure 6 depicts the diagnostic course of a patient who was initially believed to have postictal hemiparesis. Noncontrast CT revealed no abnormality in the right hemisphere of the brain. On CT angiography, however, the trunk of the middle cerebral artery could not be visualized and there were filling defects in the MCA branches. Perfusion imaging demonstrated a large deficit in cerebral blood flow and a prolongation of mean transit time, indicating that the patient was experiencing a significant degree of ischemia. Cerebral blood volume was well-preserved, however, suggesting that the brain was still viable. Blood flow values were all within normal range, with the exception of those associated with the basal ganglia. The patient was treated aggressively with a thrombolytic agent, protecting all but the basal ganglia from infarction.

CT Angiography

CT angiography represents another compelling application of neurological CT. It can be used to evaluate both the carotid arteries and the intracranial circulation for atherosclerotic disease and aneurysms.

Using a single-channel helical scanner, we image from the aortic arch to the Circle of Willis using a 3-mm collimation and a moderate pitch of 1.5. On a 4-channel CT scanner, we select a 1.25-mm collimation, the smallest collimation the detector allows. We use a pitch of 6 in the neck, as the goal is to image the carotid artery before there is significant overlap of the image from venous opacification. Once we reach the brain, timing is no longer a major issue, as the veins that opacify generally don't interfere with arterial visualization. We therefore reduce the pitch to 3 to achieve a better slice profile.

With an 8-channel scanner we are able to take advantage of the boost in speed and can improve image quality by using a moderate pitch of 10.8 in the neck. In the brain, we again use a fairly low pitch, 5 in this case, to maintain image quality and optimal slice profiles.

We always initiate angiography with enhancement-triggered scanning techniques (SmartPrep, GE Medical Systems, Milwaukee, WI). There is a latency between detection of the contrast bolus and initiation of diagnostic scanning, but the delay is shorter with modern-generation scanners. We use contrast arrival at the left heart to trigger our exams, initiating imaging from the aortic arch toward the brain as soon as we see enhancement.

Our scanning parameters change when we image the Circle of Willis alone. With a single-channel CT scanner, we use the tightest collimation, 1.0 mm, and scan at a pitch of 2 in order to get reasonable coverage. To image the Circle of Willis with a 4-channel system, we again use the thinnest possible slice, 1.25 mm, but because speed is not critical, we select a pitch of 3. With an 8-channel scanner, we take a similar approach, using the same 1.25-mm collimation but slowing the pitch to 5.

One of the greatest advantages of CT angiography over competitive techniques, including conventional and magnetic resonance angiography, is the ability to view blood vessels in relation to bony landmarks. This information is highly valuable to neurosurgeons. Figure 7 shows a nicely delineated aneurysm of the internal carotid artery, with bones in place.

Another strength of CT is the ability to do image processing on the scanner. We frequently perform CTA in acute stroke. Under such circumstances, it is important to have a very rapid and easy rendering method. We image right at the scanner, using limited-volume maximum intensity projections both obliquely through the carotids, as well as in standard orthogonal planes through the brain. This method produces images, and answers, within minutes of the patient's arrival.

Conclusion

Today, we are able to perform neurological CT with much faster scanners. Technological advances give us the opportunity, and perhaps even a mandate, to use a smaller volume of contrast media and faster injection speeds. Under most circumstances in neuroimaging, it is therefore preferable to select contrast media with the highest iodine concentration, in order to maximize opacification and enhancement while decreasing costs. Contrast media with an iodine concentration of 370 mg/mL or greater offers many advantages in neurological CT. The future availability in the United States of agents of higher iodine concentration (up to 400 mg/mL) should further extend the utility of this approach. *