 |
|
Abdominal MR angio: fast, reproducible, and safe New imaging protocols accelerate speed of abdominal MRA without sacrificing spatial resolution By Vamsi Narra, M.D.
The clinical application of three-dimensional contrast-enhanced MR angiography (3-D CEMRA) has gone mainstream.1 Major improvements in image acquisition times and reductions in the contrast volume utilized2 have made 3-D CEMRA a fast, reproducible, and patient-friendly exam. An MRA study of the abdominal aorta can be accomplished in as little as 20 minutes of scanner time. The lack of ionizing radiation and avoidance of nephrotoxicity from gadolinium-based contrast agents makes this exam safer for patients. The main factor that has driven the development and clinical use of CEMRA is the need to bypass the limitations of noncontrast-enhanced gradient-echo time-of-flight (TOF) techniques.3 Some of the drawbacks of TOF imaging are long imaging times, inability to see in-plane and small vessels, loss of signal in aneurysmal vascular structures secondary to turbulent flow, and overestimation of stenosis. For successful depiction of blood vessels based on T1 values alone, the T1 value of the blood should be considerably lower than that of surrounding tissues. The short T1 value of fat is about 270 msec, while the T1 value of blood is about 330 msec. Using a T1-shortening agent such as gadolinium-based contrast can reduce the T1 value of the blood to below 50 msec, making the vessel of interest much brighter than the background fat. In addition, the use of a T1-shortening agent allows visualization of in-plane blood vessels, as the images are no longer dependent primarily on the in-flow of the blood. This helps eliminate or reduce flow artifact. The sequences used for 3-D CEMRA are spoiled gradient-recalled-echo (SGE) sequences with zero interpolation in the partition or the slice-selected direction.4 One of the goals in optimizing the sequence parameters is to keep the overall image acquisition time shorter than 28 seconds, so that it can be achieved in one breath-hold. This goal is further aided by short TR and TE values. The short TR allows the overall reduction of the imaging time, and the short TE helps in minimizing T2* effects. Whenever feasible, the highest possible spatial resolution should be utilized. The abdominal aorta and its branches are best examined with the patient placed in the supine position. Intravenous access is obtained using a 20- or 22-gauge angiocatheter. Special care has to be taken to make sure the IV access is patent with no kinks in the tubing or occlusion of the IV line.5 Contrast can be injected manually or by using an automated MR-compatible injector. The decision whether to use a multiarray surface coil or a body coil is dependent on the extent of the coverage desired, the size of the patient, and hardware limitations, if any. In general, if a larger field-of-view is desired, it is preferable to use the body coil (Figure 1).5 If smaller coverage is needed and an improved signal-to-noise ratio is desirable, then a multiarray surface coil is preferred. Whenever possible, the patient's arms should be placed above the head or on top of the chest. When using a noncentric or sequentially ordered k-space acquisition, the center of the k-space is filled approximately three- or four-eighths into the acquisition. This is important to understand because the center of the k-space is primarily responsible for the contrast information and the periphery of the k-space is responsible for spatial resolution in the final image. Accurate timing of the peak of arterial enhancement to the center of the k-space is essential to optimizing arterial enhancement and minimizing venous contamination.6 To accurately time the peak of contrast enhancement in the abdominal vessels to the center of the k-space, the time of arrival of the contrast in the abdominal aorta must be determined first. This is accomplished by using a small test bolus. Typically, 1 to 2 cc of gadolinium-based contrast agent is injected and images are obtained in rapid succession (about one image per second) using a turbo-flash sequence. The number of images can then be counted to see when the contrast peaks in the vessel of interest so that scan delay can be calculated. Evaluation for renovascular hypertension requires additional images obtained through the kidneys and the adrenal glands to exclude any adrenal or renal pathology, which could also be the cause of hypertension (Figure 2). When the abdominal aorta is being evaluated for aneurysm, additional images are obtained to depict the outer wall of the aorta and give a true estimate of the aneurysm size. The size of the aneurysm may be underestimated on a CEMRA, as it is primarily a luminogram similar to the conventional angiogram. As the background is being suppressed on a CEMRA to depict the vessels accurately, it may be difficult to visualize the outer wall of the aneurysm. This becomes a major problem, especially when patients are being considered for endovascular stent graft placement, which requires outer-to-outer wall measurements for placement of a stent graft. The solution is to use a T1-weighted sequence or a true fast imaging with steady-state precession (FISP) sequence (Figure 3). In general, MRA images of the abdominal aorta are obtained in end inspiration. When evaluating patients with suspected mesenteric ischemia, the images should be obtained in end expiration to minimize the impression of the arcuate ligament on the celiac axis. Once postcontrast images are obtained, they are subtracted from the precontrast data set, yielding a new data set in which the background fat and other structures, which are bright on the precontrast images, are subtracted out. Images are postprocessed by maximum intensity projection (MIP).10 Although MIP images are similar to the conventional angiography images, the examiner should not draw the final conclusion based on the MIP images alone. This caution is particularly true in evaluating vessels with stenosis. An examiner using MIP images alone could overestimate the degree of stenosis. For an accurate estimate of the area of narrowing or vessel stenosis, the source images must be reviewed, and one of the best ways to do this is using a multiplanar reconstruction (MPR). All the images are reviewed using an MPR before a final conclusion can be made as to the extent and nature of the disease process involving the vascular structures. Any clinical scanner, satellite console, or PACS station or 3-D workstation can be equipped with postprocessing tools. GE scanners utilize Smart Prep, an automatic bolus detection technique. A tracker volume is placed on the vessel of interest, which in the abdomen would be the abdominal aorta. The tracker detects the baseline signal intensity levels within the vessel. Contrast is injected, and, as the baseline signal intensity goes above a defined value (approximately 30% above the base level), the MRA 3-D gradient-echo sequence is automatically triggered to acquire the contrast-enhanced MR images. This technique works very well, especially for the abdominal aorta. This technique may be less successful with smaller vessels, such as the iliacs, or with more distal vasculature. Siemens scanners use a CARE bolus technique. Two-D gradient-echo images are obtained at the rate of about one or two images per second. These images are reconstructed in real-time and displayed on the console as they are being obtained. When the contrast is seen arriving in the vessel of interest, the patient is given the breath-hold instruction and the main 3-D MRA sequence is triggered. As a rule, whenever an automatic bolus detection or similar techniques is used, a centrically reordered k-space sequence is used, because the center of k-space contributing to the contrast information on the final image should coincide with the arterial peak of the contrast. These sequences have some disadvantages, however. One of the primary disadvantages is that no phase-encoding gradient is being applied when the center of the k-space is acquired, making it very sensitive to motion. If patient motion occurs when the center of the k-space is being acquired, the entire data set will be degraded with motion artifacts. Degradation from motion artifacts is much less obvious when the patient moves during the acquisition of the periphery of the k-space. This means that the patient has to start the breath-hold before the sequence is begun when using the CARE bolus, Smart Prep, or similar techniques. Most of the 3-D MRA sequences that are used with medium resolution matrix (256) require anywhere between 22 and 28 seconds of breath-hold time. In certain clinical instances, however-when dealing with arteriovenous malformation or when a pseudoaneurysm of the vessel is suspected-it would be ideal to have a sequence with high temporal resolution. Most of the standard sequences that are available cannot accomplish this without a compromise in the spatial resolution. The speed of acquisition and the spatial resolution of 3-D CEMRA are determined and ultimately limited by the performance characteristics of the gradient and other hardware of the scanner. So far, improvement in the performance has been achieved by increasing the gradient strength, but the need to avoid neuromuscular stimulation from rapid gradient switching limits this approach. On any MR imaging system, the spatial and temporal resolutions are limited by gradient specifications for safety considerations. This limitation has inspired further development of techniques that acquire the data much more rapidly, improving the temporal resolution with no compromise in the spatial resolution and without the need for high gradient strengths. SMASH is such a technique. It uses combinations of component coil signals in a radio-frequency (RF) coil array to substitute for omitted gradient steps, reducing the burden on the gradients and allowing multiple components of the spatial encoding required to generate an MR image to be performed in parallel.12 This technique has been shown to achieve a four- to eightfold reduction in image acquisition time, with no compromise in spatial resolution. TRICKS uses an increased sampling rate for lower frequencies, temporal interpolation of k-space views, and zero-filling in the slice-encoding direction. When appropriately combined, these elements permit reconstruction of a series of 3-D image sets, having an effective temporal frame rate of one volume every two to six seconds, with no serious compromise in the spatial resolution.13 When evaluating the abdominal aorta and vasculature of the lower extremity, images were obtained using contrast injection at multiple stations. Before manual moving table devices became available, individuals started modifying their scanners so that the table could be moved in three or four stations and a single bolus chase performed. Vendors have since recognized the importance of moving tables and incorporated them into the newer scanners and software upgrades. The fundamental principle is to time the contrast in the proximal station and move the table either manually or automatically to chase the bolus down. This technique can be used not only to obtain the MR images of the lower extremity vasculature, but also in evaluating the entire aorta within the chest and abdomen, which typically involves two stations. Contrast-enhanced MR angiography has become a standard initial modality of choice at most centers for evaluating patients with renovascular hypertension. CEMRA has been shown to have very high sensitivity and specificity comparable to conventional angiograms.14 In patients who have normal renal arteries or patients with high-grade stenosis, the diagnosis can be made with certainty. MRA lags, however, in patients with a moderate amount of stenosis or in patients with moderate stenosis in whom a determination must be made as to whether the stenosis is greater or less than 50%. This information is critical, as patients with hemodynamically significant stenosis (>50%) need intervention, and MR angiography cannot always estimate the degree of stenosis accurately and consistently. Prince et al have introduced the technique of phase-contrast images based on the principle that if serious stenosis of a renal artery is present, this region will show loss of signal due to turbulence and phase shift of the spins. This technique has been useful in evaluating and triaging patients with mild to moderate stenosis.15 Volumetric interpolated breath-hold examination (3-D VIBE) imaging was initially described by Rofsky et al.16 This sequence is also a 3-D gradient-recalled-echo sequence similar to that of a 3-D sequence used for CEMRA, with two main differences. The echo read-out in the KY direction for a VIBE sequence is symmetric, in contrast to an asymmetric echo for a dedicated MRA sequence. In addition, the flip angle is reduced to 12° to 15°. The purpose of these two changes is to improve the signal from the background tissue structures. This technique can be applied in the evaluation of the abdomen and pelvis and used in evaluating the solid organs while simultaneously acquiring vascular information (Figure 4). We will soon see more applications like SMASH and TRICKS, which will be based on the principle of acquiring data rapidly and gaining a temporal resolution that does not compromise spatial resolution. With improvements in image processing time, it is not impossible that a combination of these rapid image acquisition techniques and faster computer postprocessing could make higher image matrix (512) routine in CEMRA. Intravascular contrast agents aimed at increasing the duration of the time the agent stays in the vascular space are undergoing phase II and phase III clinical trials. Initial data from these studies are very promising, as high spatial resolution images can be obtained as late as 50 minutes following administration of a single dose of contrast. This introduces a new problem of venous contamination, however, and investigators are actively working on effective automated venous subtraction techniques. If MRA is to largely replace conventional angiography, it has to be taken a step further. Detecting the stenosis is no longer enough; MRA must be able to accurately estimate the severity of the stenosis and help triage these patients for conservative management or intervention. Different techniques are being tried, specifically for renal artery stenosis or iliac stenosis. These include velocity measurements and indirect calculation of the pressure gradient across an area of stenosis. This area in particular will see tremendous growth in the next few years. References
|
CancerNetwork | ConsultantLive | Diagnostic Imaging | Psychiatric Times | Search Medica | Physicians Practice
© 1996 - 2009 CMPMedica LLC, a United Business Media company
Privacy Statement - Your California Privacy rights - Terms of Service