Applied Radiology

EDITOR'S NOTE

As MR imaging technology has advanced, so has accurate depiction of the carotid arteries. With the stronger, faster gradients typical of echo planar imaging (EPI) MR systems, contrast-enhanced MR angiography (CE-MRA) can now be performed during the initial passage of gadolinium through the arterial system, effectively eliminating venous contamination.

CE-MRA allows a much more accurate depiction of carotid stenosis and ulceration than unenhanced MRA techniques. Clinical applications and the technical details behind this new technique are discussed in this issue of Applied Imaging.­­William G. Bradley, Jr., MD, PhD, FACR

Contrast-enhanced MR Angiography of the Carotids

The term contrast-enhanced MR angiography (CE-MRA) generally refers to the rapid acquisition of vascular images while a bolus of gadolinium passes through the arterial phase following an intravenous injection. Typically, CE-MRA requires the strong, fast gradients that are used for echo planar imaging (EPI). Thus, while gadolinium has been used occasionally to supplement three-dimensional (3D) time-of-flight (TOF) MRA over the years, eg, for demonstration of very slow flow or the relationship of an enhancing tumor to blood vessels, CE-MRA has really only come into its own since EPI units were installed on a widespread basis 4 to 5 years ago.

At the heart of CE-MRA is the effect of saturation on flow-related enhancement. ¹ When fully magnetized, ie, unsaturated, spins enter the first slice of an imaging volume, and return a strong signal after exposure to the first radiofrequency (RF) pulse ("flow-related enhancement"). ^1,2 For thin, two-dimensional (2D) TOF slices oriented perpendicular to the direction of blood flow, spins are unlikely to be excited by more than one RF pulse. For a thicker 3D TOF slab, the spins may be exposed to multiple RF pulses and eventually run out of magnetization (Figure 1). At this point, they are said to be "saturated." The effect of saturation is magnified when the spins flow parallel to, ie, within, the slice rather than perpendicular to it. This is because they are now susceptible to multiple RF exposures over the greater distance of the field-of-view (FOV), eg, 25 to 35 cm. Gadolinium shortens the T1 of the flowing protons, allowing them to sustain multiple RF exposures without saturation, even while traversing the large distances of a typical FOV (Figure 2). Thus, CE-MRA is generally performed in the same (coronal) plane as the arteries . ²

CE-MRA typically involves a bolus injection of gadolinium by a power injector with acquisition of a 3D TOF slab. Inplane spatial resolution is typically on the order of 1 mm, while slice thickness is on the order of 1 to 2 mm. Typically, zero interpolation (ZIP) is applied in the slice direction to minimize "stair-step" artifacts when the images are subsequently computer-processed by a maximum intensity projection (MIP) and rotated about a vertical axis to be viewed from any angle. ³ Since the base sequence for a 3D TOF application is a T ₁ -weighted gradient echo, subcutaneous fat normally appears bright. For this reason, we typically acquire a pre-injection mask image and then subtract it from the image acquired during the arterial phase.

CE-MRA is typically used to image arteries such as the carotids, which places certain technical demands on the gradients. Specifically, they need to be able to complete an arterial phase acquisition before the gadolinium gets to the venous phase­­a period of approximately 10 sec (Figure 3). There are two ways to acquire the data to ensure minimal venous opacification. The first technique consists of multiple 10-sec acquisitions following the injection of contrast­­one of which will predominately cover the arterial phase. This "multiphase" acquisition typically consists of a 10-sec mask image (Figure 4A) followed by 4 or 5 separate 10-sec acquisitions (Figures 4B and C). The arterial-phase image is then selected, masked, and subjected to the MIP algorithm for viewing from any projection (Figure 4D). ^4,5

More recently, a clever scheme for data acquisition has been introduced that allows higher resolution imaging of arteries while still suppressing the signal from veins. This technique is known as "elliptical-centric" CE-MRA, referring to the filling of k-space in two dimensions. ⁶

[k-space is a mathematical construct that facilitates visualization of different fast MR imaging techniques. ^7,8 Diagrammatically, k-space consists of a matrix of 256 ¥ 256 points (ie, the number of frequency and phase points), with each phase value corresponding to a horizontal row and each frequency sampling point corresponding to a vertical column. Typically, the most negative values of the phase-encode gradient are at the bottom of the matrix, those with the weakest phase encoding are in the center, and the k-space lines with the strongest value of the phase-encode gradient are at the top. Since weaker phase encoding leads to less dephasing, most of the signal comes from the center of k-space. The idea behind elliptical-centric k-space acquisition is to start acquiring in the center of k-space just as the contrast hits the arterial phase. The term "centric" k-space coverage refers to a single-slice technique (eg, 2D-TOF), while the term "elliptical-centric" refers to a 3D-TOF technique in which there are two phase directions (ie, traditional phase and slice).]

In order to ensure that the acquisition does not begin until the contrast begins to fill the arteries, either a timing run must be performed or automated bolus detection software must be used. Such software (ie, SmartPrep [GE Medical Systems, Milwaukee, WI] BolusTrak [Philips Medical Systems, Bothell, WA] and CareBolus [Siemens Medical Systems, Iselin, NJ]) is particularly useful for the abdominal aorta and below; however, it is less reliable for imaging the carotid arteries. Therefore, a timing run is usually performed, which consists of a 2-mL injection of gadolinium with image acquisition every second or so at the level of the carotid bulb. While many schemes have been utilized to optimally determine the timing delay, we simply wait until the carotid bulb has first reached its maximum intensity and use that as the timing delay.

One of the advantages of elliptical-centric k-space coverage is that a much longer time is available for acquisition, eg, 30 to 60 sec, compared with the 10-sec multiphase acquisition. This allows higher spatial resolution while still suppressing venous signal, since it is acquired in the low signal-to-noise periphery of k-space (Figure 5). Our current technique is spoiled gradient echo (SPGR) 6.3/1.5/45š (with a 224 ¥ 256 matrix acquired over a 26-cm FOV. We use a 31.25 kHz bandwidth, a single excitation, and a partial Fourier acquisition (80%) in the first-phase direction. The second-phase direction has 50 1.2-mm thick slices. We zero interpolate in the slice direction and to 512 ¥ 512 in plane.

We typically start a 20-gauge intravenous line in an antecubital vein and connect it to a power injector. A 2-mL bolus of gadolinium is injected to determine the timing delay. Then, following acquisition of a mask image, the injection is begun and, following the predetermined timing delay, the acquisition of contrast-enhanced images is performed. Following substraction, the data is MIPPed and displayed every 15š of rotation about a vertical axis.

Clinical Utility

Contrast-enhanced MRA is more accurate in quantifying stenoses than conventional, unenhanced 2D TOF or 3D TOF MRA. While unenhanced techniques are prone to saturation effects that overestimate the degree of stenosis, the gadolinium in a CE-MRA technique effectively eliminates saturation effects, resulting in a truer rendering of the degree of stenosis. Even "string signs," ie, 99% occlusions (Figure 6), can be detected with CE-MRA. Since the vertebral arteries are also included in the coronal slab, vertebral artery stenosis (Figure 7) can also be detected. Carotid bulb irregularities suggestive of ulceration can be detected easily with CE-MRA (Figure 8).

Intracranial Vascular Screening

Using conventional volume neck coils, the origins of the carotid and vertebral arteries are routinely visualized on a CE-MRA examination (Figures 4D and 5). By expanding the FOV to approximately 30 cm in the read direction (along the long axis of the body), the acquisition can be extended through the circle of Willis without venous contamination (Figure 9). While the intracranial vessels are thereby displayed with lower resolution than our usual unenhanced multiple overlapping thin-slab acquisition (MOTSA) MRA brain study, ⁹ it still serves as a good screening study of the proximal vessels in the Circle of Willis. While CE-MRA can be performed in the brain, the necessity for a longer acquisition to provide higher resolution typically results in greater venous contamination; therefore, gadolinium is rarely used in the brain at present. Perhaps with venous editing techniques in the future, CE-MRA will be applied to the brain as well.]

Conclusion

As MR technology continues to advance, so does accurate, noninvasive depiction of the carotid arteries. With the stronger, faster gradients present on echo planar MR systems, CE-MRA can now be performed much more accurately for sizing stenoses than was possible with previous unenhanced MRA techniques. In many centers, the combination of a positive contrast-enhanced MRA and a positive duplex Doppler ultrasound is a sufficient indication for a carotid endarterectomy, obviating the need for the more expensive, and more invasive, catheter angiogram.

References

1. Bradley WG, Waluch V. Blood flow: Magnetic resonance imaging. Radiology . 1985;154: 443-450.

2. Bradley WG. Flow Phenomena. In: Stark DD, Bradley WG (eds). Magnetic Resonance Imaging, 3rd ed. St. Louis: Mosby, Inc. 1999:231-256.

3. Masaryk TJ, Perl J II, Dagirmanjiam A, et al. Magnetic resonance angiography: Neuroradiological applications. In: Stark DD, Bradley WG (eds). Magnetic Resonance Imaging, 3rd ed. St. Louis: Mosby, Inc. 1999:1277-1316.

4. Parker DL, Goodrich KC, Alexander AL, et al. Optimized visualization of vessels in contrast enhanced intracranial MR angiography. Magn Reson Med . 1998;40:873-882.

5. Erly WK, Zaetta J, Borders GT, et al. Gadopentetate dimeglumine as a contrast agent in common carotid arteriography. AJNR Am J Neuroradiol. 2000;2:964-967.

6. Fain SB, Riederer SJ, Bernstein MA, Huston J 3rd. Theoretical limits of spatial resolution in elliptical-centric contrast enhanced 3D MRA. Magn Reson Med. 1999;42:1106-1116.

7. Bradley WG, Chen D-Y, Atkinson DJ. Fast spin echo. In: Bradley WG, Bydder GM (eds). Advanced MR Imaging Techniques. London: Dunitz, 1997:3-29.

8. Bradley WG, Atkinson DJ, Chen D-Y. Using high performance gradients. In: Bradley WG, Bydder GM (eds). Advanced MR Imaging Techniques. London: Dunitz, 1997:31-62.

9. Melhem ER, Poon EK, Weinreich DM, et al. Comparison of 2D- and 3DFT multiple overlapping thin-slab acquisition TOF MR angiography in carotid disease. J Neuroimaging . 1998;8(1):3-7.

Clinical Quiz: True or False

1. Contrast-enhanced MRA (CE-MRA) should be performed in every patient with stroke or TIA.

2. CE-MRA requires an arterial injection.

3. CE-MRA can be performed on any modern MR system.

4. CE-MRA can only be performed on high-field systems.

5. CE-MRA can be performed in more than 60 seconds and still minimize the signal coming from veins.

1. False. Patients with cardiac pacemakers and ferromagnetic intracranial aneurysm clips cannot undergo MRA.

2. False. Contrast-enhanced MRA is performed with an intravenous injection that catches the gadolinium bolus during the arterial phase.

3. False. Only those systems with high-performance gradients, eg, echo planar systems and cardiovascular systems, can perform CE MRA.

4. False. Low-field systems with fast, strong gradients can perform CE MRA as well.

5. True. Using elliptical centric techniques, the examination can be extended to 60 seconds, improving the spatial resolution while minimizing the signal from enhancing veins. This is accomplished by filling the high signal-to-noise center of k-space during the arterial phase and the low signal-to-noise periphery of k-space during the venous phase.

Note: No contrast agents are approved by the U.S. Food and Drug Administration for use in magnetic resonance angiography.

EDITORIAL STAFF

William G. Bradley, Jr., MD, PhD, FACR * Editor-in-Chief

O. Oliver Anderson * Publisher

Elizabeth A. McDonald * Editor

Beverly Harris * Assistant Editor

Felice Ponger-Shaloum * Art Director/Production Manager

Applied Imaging is published by Anderson Publishing, Ltd. It is supported by a grant from Amersham Health. The views and opinions expressed in this publication are those of the authors and do not necessarily reflect those of the publisher or sponsor. All inquiries should be addressed to: Anderson Publishing, Ltd., 1301 W. Park Avenue, Ocean, NJ 07712.