Technique

The typical cervical MRA examination should encompass the top of the aortic arch, inferiorly, and the skull base, superiorly. The extracranial course of both carotid and both vertebral arteries should be included in this field-of-view (FOV). The radio frequency (RF) imaging coil should ensure adequate visualization of this large area of coverage while still maintaining a high signal-to-noise (SNR) ratio, which is required for proper visualization of vessel detail. Dedicated phased array neu-rovascular RF coils are available from a number of manufacturers and are the preferred imaging coil design for this purpose. These coils allow head/ brain examinations that are comparable in quality to dedicated head coil MR imaging exams while, at the same time, also provide high SNR images of the neck and thoracic inlet, including the top of the aortic arch. These neurovascular RF coils allow extracranial MRA, intracranial MRA and brain MRI exams to be acquired during the same MR study without the need to move the patient or to change coils during the examination.

A number of different pulse sequences are available for cervical MRA imaging. Each has its own advantages and disadvantages. The 2D TOF technique using an inferior "walking" saturation pulse that moves with each serially acquired slice was originally devised and popularized by Keller, et al [6]. This technique has the advantage that it can encompass any length of coverage that is needed by merely prescribing an adequate number of slices. Axial 2D TOF acquisitions can provide fairly uniform vessel visualization throughout the neck since each 2D slice is acquired separately. The use of thin slices will minimize flow saturation effects if blood flow is brisk. The in-plane resolution on axial 2D TOF MRA is very good; but the longitudinal resolution is limited to about 1.5 - 2 mm due to the limitations imposed by the minimum slice thickness the MR scanner can generate together with the traditional practical consideration of imaging time. The disadvantages of this technique include a limited SNR due to the fact that each slice is acquired as a separate thin 2D acquisition rather than averaging the entire volume of slices in a 3D acquisition. This technique also has the disadvantage of overestimating the degree of stenosis when higher grade vessel narrowing and/or slow flow is present [7, 8]. In addition, 2D TOF technique is highly sensitive to even small amounts of patient motion that can lead to the so-called "stairstep" artifact due to misregistration of adjacent slices. This artifact can be seen even with small amounts of motion that may be produced through normal quiet breathing and certainly is heightened by swallowing (Fig. 10). An additional limitation of this technique is the poor temporal resolution, which can take as long as 8-12 minutes to acquire the total number of slices needed, which usually number 80100 slices of 1.5mm thickness. However, despite these limitations, this technique remains a viable alternative for cervical vessel MRA.

Another major technique is 3D TOF MRA. This can be acquired either as a single 3D volume of images or as a series of multiple overlapping 3D volumes, which can be combined to provide a single longitudinal segment of coverage. In either case, the major limitation of this technique is that it generally cannot be used to visualize the entire length of coverage needed for a complete evaluation of the cervical vasculature. Thus, this technique is most often used as a supplemental imaging technique to evaluate the common carotid artery bifurcations in more detail or to better estimate the degree of stenosis. Unlike 2D TOF, the 3D TOF technique is much more robust for accurately determining the degree of stenosis [7,9,10]. These targeted 3D TOF sequences also generally require less time to image compared with the 2D TOF MRA of the entire neck. They are also less affected by minor degrees of motion such as that caused by quiet breathing although they are significantly degraded by major degrees of motion such as swallowing or gross patient movement of the head and neck.

Contrast enhanced 3D MRA (CE 3D MRA) is the preferred method (Fig. 11) for total evaluation of the extracranial cervical vasculature [11-13]. This method utilizes a rapid T1-weighted fast 3D gradient echo pulse sequence. Imaging is typically performed in the coronal plane during the first pass of contrast following an intravenous bolus injection of a gadolinium-chelate contrast agent. Acquisition time for this sequence is generally between twenty and forty seconds and this rapid acquisition time results in far fewer motion artifacts when compared with either traditional 2D TOF or 3D TOF imaging sequences. These rapid imaging sequences are best acquired with elliptical centric k-space phase encoding that allows the low spatial frequency data (i.e. center of k-space) for each of the phase encoding axes to be acquired at the beginning of the pulse sequence so that maximal contrast weighting occurs early in the image acquisition. This results in both increased resistance to breathing artifacts and a good separation of arteries from veins even with acquisition times of forty seconds or longer [14].

Another emerging technique utilizes time-resolved imaging whereby only a portion of the k-space phase encoding steps are selectively repeated with each subsequent temporal phase. 3D TRICKS (Time-Resolved Imaging of Contrast Kinetics) is one such technique [15] and it utilizes the periodic refreshing of the high contrast (lower order)

Artifact Mra Cervical

Fig. 10a-c. 2D TOF MRA of the neck in a normal individual. a The normal neck vessels are visualized from the level of the aortic arch to just below the skull base. Note the normal appearance of the common carotid artery bifurcations on both sides. The vertebral arteries are also well shown. Also note the marked "stairstep" artifacts present due to slight patient motion between individual axial source images. This artifact is most pronounced within the lower portion of the MRA at the level of the aortic arch and origin of the great vessels. Combination of normal pulsation motion and breathing motion causes severe stairsteping with resultant loss of signal and detail. b Axial source images at the level of the common carotid artery bifurcations show a flow artifact along the posterior wall of both internal carotid artery bulbs (arrows). This is due to signal loss from turbulence and high velocity flow. This type of artifact is accentuated with higher degrees of stenosis and can lead to overestimation of stenosis. c Axial source images a few millimeters above B now shows that the vessel outlines are well demonstrated in both internal and external carotid branches as well as in both vertebral arteries

High Velcoity The Vertebral Artery

Fig. 11. CE MRA in a normal individual. Coronal CE MRA obtained during the first pass of a bolus of contrast material shows visualization of the neck vessels from the aortic arch to the skull base. There is good visualization of the great vessel origins as well as the carotid and vertebral arteries. This is the same patient as illustrated in Fig. 10. Note the greatly improved appearance of the aortic arch and great vessel origins. Note also the uniformly smooth vessel outlines without stairstep artifact. It is for this improved visualization that this is the preferred neck MRA technique phase encoding steps interleaved with higher order phase steps that are combined for high spatial resolution images with high temporal resolution. Elements of this method include an increased sampling rate for centric k-space lines, temporal interpolation of k-space views, and zero filled interpolation in the slice-encoding dimension. By sharing vital k-space data, this technique results in the generation of multiple high resolution 3D data sets representing 4-9 second time points for dynamic visualization of arterial wash-in and venous wash-out of the contrast bolus.

In order to obtain high quality images with arterial-phase CE MRA the timing of the contrast bolus and the rate of injection is crucial [16]. The period of preferential carotid enhancement is typically brief (e.g. as short as 5 seconds). Imaging too late can result in significant jugular venous contamination of the images and poor carotid visualization. There are a number of options for contrast bolus timing with the start of the image acquisition that are available with different manufacturers and MR scanner models. The simplest technique and one that is very reliable is to deliver a small test injection of contrast (usually a 2 cc bolus) while serially imaging a fixed region of interest in the neck, generally the common carotid artery just below the bifurcation [17]. Using this method a time versus contrast enhancement curve is plotted that shows the time-of-arrival post injection of peak contrast enhancement. This time delay is then used to program the start of the CE MRA acquisition following the start of injection of the contrast bolus. Since the time delay following start of injection and the rate of contrast injection are both critical for obtaining good quality CE MRA that is reproducible across different subjects, an automated contrast injection using an MR compatible injector is essential.

An alternate method for triggering the start of the CE MRA sequence utilizes automated detection of contrast arrival [18]. This method generally involves placing a region-of-interest (ROI) over a vessel on the scout image in which signal intensity is sampled on rapid serially acquired measurements. Pre-determined thresholds are set for an amount of signal intensity change within this preselected monitoring ROI and when this threshold is detected the start of the CE MRA sequence is automatically triggered. While this method may seem useful and time efficient by allowing one to bypass the need to acquire and analyze a test bolus injection, in the cervical region at least, this method is has several pitfalls. The reasons for this are several fold and include improperly set thresholds, variations in the degree of contrast enhancement among different patients and, possibly most importantly, sporadic detection errors caused by partial volume effects due to the small size of the neck vessels relative to the chosen ROI. If the ROI is too small, low SNR in the monitoring volume or slight patient motion may limit detection of small amounts of enhancement. On the other hand, if the sampling volume is too large, averaging of the vessel with surrounding non-enhancing tissues may also yield poor contrast material detection. In either case, the result may be unreliable detection of contrast bolus arrival. In our experience, automated triggering has proved too unreliable for cervical vessel MRA and we have discarded its use in favor of test injections.

More recently, MR fluoroscopic triggering has been used for carotid CE MRA [19]. This method uses serial rapidly acquired and instantaneously displayed low-resolution 2D images (i.e. MR fluoroscopy) of the area of interest to detect the time of arrival of the full contrast bolus. When the contrast bolus is visually detected on the serial images, the sequence is then triggered manually by the observer. This technique, although available only on newer MR scanners, is the most rapid and reliable technique.

As already indicated, an automated MR-compatible contrast injector is a very important piece of equipment if one is to obtain optimized and reproducible carotid CE MRA. Timing of the injection relative to triggering of the start of the CE MRA sequence is important, as is an accurate and reproducible rate of contrast media injection. The optimal rate of contrast media injection has been shown to be approximately 2 mL per second [16]. If the contrast bolus is injected too rapidly it can result in signal loss from T2* effects and an increased blurring of vessel outlines due to the short bolus duration of gadolinium which does not cover a large segment of k-space when injected too rapidly [16]. These artifacts are generally seen at contrast injection rates in excess of 4 mL per second. In addition high injection rates may result in a retrograde filling of the jugular vein, which than causes venous overlay on arterial phase images with poor carotid visualization.

The total dose of contrast used for neck MRA is a full 20 mL vial of Gd-chelate contrast agent (minus the 2 mL used for a test injection if needed). Injection of this 18-20 mL bolus of contrast material is followed immediately by a 20 mL bolus of normal saline injected at the same rate (2 mL/sec) in order to flush the contrast agent rapidly through the arm veins and superior vena cava. This helps to ensure that the full dose of contrast reaches the cervical vessels in a uniform bolus.

Mistiming of a contrast bolus can result in poor MRA images. If the imaging starts too early relative to bolus arrival then there is poor vessel contrast since the high contrast-weighted phase encodings are acquired early in the elliptical centric ordered image acquisition. On the other hand, if

Fig. 12. CE MRA demonstrates overlap between arterial and venous phases. The acquisition was mistimed and started a few seconds after arrival of the contrast bolus. There is simultaneous visualization of both arteries and veins filled with contrast. This results in obscuring of arterial vessel details and greatly interferes with diagnostic interpretation

Vasulature The Neck Mra

the image acquisition is delayed relative to arrival of a contrast bolus in the neck vessels, this will result in overlap between arteries and veins making diagnostic interpretation of the CE MRA difficult and less sensitive (Fig. 12).

Post processing of the images is done following acquisition of the MRA. This is done whether the sequences are acquired as 2D TOF, 3D TOF or CE MRA. Maximum intensity projection (MIP) images are obtained of the entire volume of images. These can generally be performed automatically by the MR scanner as part of its standard image post-processing , whereby a series of longitudinal radial projections of the neck vessels are projected from multiple different angles of rotation around the neck (Fig. 13a-c). In addition, we also generally acquire targeted MIP images where the ipsilater-al carotid and vertebral arteries are isolated from

Fig. 12. CE MRA demonstrates overlap between arterial and venous phases. The acquisition was mistimed and started a few seconds after arrival of the contrast bolus. There is simultaneous visualization of both arteries and veins filled with contrast. This results in obscuring of arterial vessel details and greatly interferes with diagnostic interpretation

Fig. 13a-c. Normal contrast enhanced 3D MRA. a-c Direct coronal view of the first pass 3D MRA acquired with elliptical centric phase encoding technique. Note good visualization of all of the major vessels as well as many smaller vessels within the neck. Also note good separation of arteries and veins, the latter not being visible. band cThe 3D MRA can be rotated in different projections as illustrated

Mra Neck Anatomy

Fig. 15a, b. Edited MIP projection of the left carotid and left vertebral arteries obtained from a CE MRA. Two representative projections illustrate good demonstration of the left common carotid artery bifurcation, which is normal in appearance as well as a small left vertebral artery. Note irregular narrowing at several points in the left vertebral artery caused by extrinsic compression from vertebral osteophytes. Also note good demonstration of the origin of the left vertebral artery using the edited MIP technique. a Vessels demonstrated in LAO projection. b Vessels demonstrated in RAO projection

Fig. 14. CE 3D MRA can also be edited such that only selected portions of the vessel system are displayed using a targeted MIP algorithm. Note here that this view shows only the left carotid and left vertebral arteries while other vessels have been selectively removed in order to improve visualization without overlap and interference from other vessels

Fig. 15a, b. Edited MIP projection of the left carotid and left vertebral arteries obtained from a CE MRA. Two representative projections illustrate good demonstration of the left common carotid artery bifurcation, which is normal in appearance as well as a small left vertebral artery. Note irregular narrowing at several points in the left vertebral artery caused by extrinsic compression from vertebral osteophytes. Also note good demonstration of the origin of the left vertebral artery using the edited MIP technique. a Vessels demonstrated in LAO projection. b Vessels demonstrated in RAO projection

Fig. 14. CE 3D MRA can also be edited such that only selected portions of the vessel system are displayed using a targeted MIP algorithm. Note here that this view shows only the left carotid and left vertebral arteries while other vessels have been selectively removed in order to improve visualization without overlap and interference from other vessels the contralateral vessels using a targeted region of interest (Fig. 14). By selecting or editing the region of interest the MIP algorithm is not only restricted to the vessels of one side of the neck but also much of the anatomical structures and high signal intensity subcutaneous fat is excluded as well. These edited MIP images (also called sub-volume MIPs) of isolated vessels render a much higher quality MRA image that is unencumbered by overlapping vessels from the opposite side or by artifacts caused by high intensity background structures such as bone marrow fat or subcutaneous fat (Fig. 15a, b). The latter may result in additional noise on the full FOV MIP images. The targeted MIP images can be done either on the MR scanner operator console or using an offline independent computer workstation where generally there are additional tools that allow one to color the vessels and to introduce 3D shading to provide an improved visual impression of vessel anatomy.

The MIP images provide angiographic-like views that readily display anatomical configuration and location, and are thus easy to interpret. However, it is also imperative to review the individual source images that make up the entire MRA volume. This is important since the maximum intensity projection algorithm sets arbitrary thresholds for including or rejecting pixels. Thus, some areas of shading that actually represent the vessel lumen may be excluded on the MIP projection while other high signal intensity non-vascular structures (i.e. fat, hemorrhage, etc.) may be projected onto the vessel outline and appear as part of the vessel. In either case, this can result in misdiag-nosis. These potential pitfalls can be avoided and the true vessel lumen better appreciated when visualizing the actual cross sectional source images. This is especially critical when evaluating the 2D TOF and 3D TOF images. It appears to be less crucial, although still important and advisable, when evaluating the CE MRA images. The latter is less critical probably due to the fact that the source images for the CE MRA are acquired in coronal projection, parallel to the vessel of interest rather than presenting cross sectional pictures of the vessel of interest.

Was this article helpful?

0 0
Essentials of Human Physiology

Essentials of Human Physiology

This ebook provides an introductory explanation of the workings of the human body, with an effort to draw connections between the body systems and explain their interdependencies. A framework for the book is homeostasis and how the body maintains balance within each system. This is intended as a first introduction to physiology for a college-level course.

Get My Free Ebook


Responses

  • keri
    Can swallowing during a neck mra create artifact?
    8 years ago

Post a comment