 |
|
|
The Office of Health Technology Assessment (OHTA) evaluates the risks, benefits, and clinical effectiveness of new or unestablished medical technologies. In most instances, assessments address technologies that are being reviewed for purposes of coverage by federally funded health programs.
OHTA's assessment process includes a comprehensive review of the medical literature and emphasizes broad and open participation from within and outside the Federal Government. A range of expert advice is obtained by widely publicizing the plans for conducting the assessment through publication of an announcement in the Federal Register and solicitation of input from Federal agencies, medical specialty societies, insurers, and manufacturers. The involvement of these experts helps ensure inclusion of the experienced and varying viewpoints needed to round out the data derived from individual scientific studies in the medical literature.
After OHTA receives information from experts and the scientific literature, the results are analyzed and synthesized into an assessment report. Each report represents a detailed analysis of the risks, clinical effectiveness, and uses of new or unestablished medical technologies. If an assessment has been prepared to form the basis for a coverage decision by a federally financed health care program, it serves as the Public Health Service's recommendation to that program and is disseminated widely.
OHTA is one component of the Agency for Health Care Policy and Research (AHCPR), Public Health Service, Department of Health and Human Services.
Thomas V. Holohan, M.D.
Director
Office of Health Technology Assessment
Clifton R. Gaus, Sc.D.
Administrator
Agency for Health Care Policy and Research
Questions regarding this assessment should be directed to:
Office of Health Technology Assessment
AHCPR
Willco Building, Suite 309
6000 Executive Boulevard
Rockville, MD 20852
Telephone: (301) 594-4023
Magnetic resonance angiographies (MRAs) are magnetic resonance imaging (MRI) techniques applied to the noninvasive visualization of both blood flow and blood vessel morphology in which images are created in a projection format similar to that of conventional x-ray angiography (CA). MRA has evolved as a modification of MRI technology, in which the body is placed in a homogeneous magnetic field and a selected volume of flowing blood is magnetically depolarized with a radio frequency (RF) pulse to act as an internal contrast agent to static tissue which can generate images using computer hardware and software.(1,2) Contrast between flowing blood and surrounding tissue can be obtained because the excited protons of the flowing blood in the RF fields or magnetic-field gradients move out of the imaging field before signal acquisition by the sensor, resulting in a differential high-signal intensity in the blood vessels compared with the magnetically saturated surrounding tissues.(3) MRA signals are MRI signals encoded with spatial information by one of a variety of imaging techniques, , Fourier imaging, echo-planar imaging, projection reconstruction, and spiral scanning.(4) Ideally, MRA achieves adequate image resolution with freedom from flow-related artifacts during a short imaging time.(5) MRA is predominantly accomplished by fast imaging methods, which permit the accrual of more information in a shorter time than do conventional scans, e.g., 2-3 minutes vs. 20-30 minutes.(4)
MRA creates a large number of thin, contiguous-image slices that can be viewed individually or reconstructed into a slab of images collapsed into a three-dimensional (3D) projection for visualization of both the lumen and the surrounding vessel wall.(6) Projected 3D images can thus be regarded as a "flow map" of the vasculature which incorporates both anatomic and physiologic information.(7)
The primary issues addressed in this assessment are the safety, effectiveness, and clinical utility of MRA for the determination of blood vessel morphology and the measurement of blood flow. In addition, indications, contraindications, and patient selection criteria will be addressed.
All materials exhibit some magnetic properties caused by electron motion and nuclear spin.(2) Proton MRI technology is based on the principle that protons within body tissues and fluids can absorb and subsequently emit RF signals when placed in a magnetic field.(8),[a] These signals can be detected and displayed to provide data for spatial location and contrast discrimination of specific tissues based on their individual biophysical characteristics. Signal changes during MRI occur as the result of blood flow, i.e., moving protons through magnetic-field gradients and RF fields. MRA uses these changes to differentiate flowing blood from surrounding stationary structures and generate images (angiograms) depicting vasculature based on flow phenomena.
The application of MRI principles to the noninvasive measurement of blood flow in humans began at the National Institutes of Health (NIH) in 1956 and was first reported in 1959.(9) These blood-flow measurements depended on the interaction of motion with RF excitation and involved spin-tagging experiments in which a bolus of magnetically "tagged" spins (protons) moving along a vessel were subsequently detected downstream.(10, 11) In extending the techniques from tagging experiments to angiography, vessel morphology was extracted and a series of contiguous sections were reformatted using computer processing. Determination of velocity can also be derived by measuring changes in signal amplitude between the stronger signal of freshly magnetized spins compared with that of saturated spins.(9, 12) Problems of poor spatial resolution, which typically occur when reformatting angiography images, were later circumvented by using phase contrast and subtraction to produce direct images of pulsatile arterial flow.(13)
The earliest clinical application of MRA was described by Weeden et al in 1985.(14). Their technique exploited those differences between systolic and diastolic arterial flow on cardiac-gated spin-echo images. In that same year, Hale et al(15) described the first MRA 3D reconstruction of blood vessels.
The subsequent development of additional MRA techniques and their successful clinical application have been derived primarily from the variety of methods for image acquisition and complementary display algorithms.(16) MRA methods manipulate variables in RF, including timing and rates of change, to generate signals from flowing blood, while signals from stationary tissue are subtracted or suppressed in a manner analogous to that used with conventional digital subtraction angiography.(17-20) In effect, MRA angiograms are physiologic studies (rather than anatomically weighted conventional angiograms) in which moving blood, as the contrast agent, is displayed as high-signal intensity.(21, 22) The nature of the magnetic resonance (MR) signal from flowing blood is complex and partially depends on the spin density and relaxation properties of the blood plus the dynamics of blood flow and the thickness of the imaged section.(23)
The effects of flow in MRA include signal enhancement caused by the inflow of spins before exposure to a RF pulse (unsaturated), signal loss caused by outflow of excited spins from the imaged section, and dephasing of signal caused by motion along magnetic-field gradients.(24) Signal intensity is determined by the combined effects of these mechanisms.
The goal of MRA in imaging vessels is to produce a signal from the blood that has sufficient contrast to permit an image of the vasculature to be extracted from the background tissue in adequate detail.(25) There are a variety of MRA techniques used to accomplish this, and the method selected is related to its clinical application.(26) The techniques applied to isolate the vasculature from surrounding tissue, and the differential signal enhancement for fast and slow flow, form the basis for vascular imaging obtained during systolic or diastolic flow.(27) These techniques, in effect, render the background structures transparent and enable the 3D vasculature to be projected and imaged on a two-dimensional (2D) image plane.(14) Using MRA technology, a single vessel may be viewed separately by limiting the field of view of the postprocessing algorithm used to display individual slices from volume-image data.(22)
Images must be acquired during diastole for arterial visualization. Specialized pulsing sequences, not generally used for routine clinical imaging, are required for the determination of blood-flow velocity and direction.(28) None of these techniques requires the use of contrast media (although some may be enhanced by them).(29-32)
All MR flow measurements and imaging techniques exploit either longitudinal (tagging or wash-in-wash-out procedures for the inflow or outflow of saturated or unsaturated spins) or transverse spin magnetization (phase-shift or dephasing procedures, which compensate for a decreased MR signal) (33, 34) These commonly preferred MRA techniques are termed time-of-flight (TOF) or phase-contrast (PC) methods.(5, 27, 35) TOF effects refer to signal changes generated by protons flowing perpendicular to the plane of sectioning during a MRI pulse sequence and are related to the increase in amplitude of the signal returning from flowing blood in the imaged area. TOF effects occur whenever blood flows out of a region in which its longitudinal magnetization has been modified. PC effects relate to the increase in phase angle of the returning signal (which is proportional to the velocity of flowing blood) and occur whenever blood having transverse magnetization moves in the direction of a magnetic-field gradient.(7, 26)
TOF methods, first described by Feinberg et al in 1984, (36) use a velocity-compensated sequence, which refocuses moving MR spins to obtain the proper contrast based on the differential saturation of magnetization between flowing blood and surrounding tissue.(37) Imaging MRA using TOF relies on flow-compensated gradient-echo pulse sequences. Using this technique, nonsaturated spins of blood flowing into an imaged volume produce a strong signal (flow-related enhancement) and appear bright when a motion-artifact-suppression technique (gradient-moment nulling) is used, and black when presaturation is used.(8, 38-40) Gradient-moment-nulling techniques compensate for velocity, acceleration, and motion errors by modifying gradient structures, whereas presaturation uses a set of RF pulses to saturate blood flow or moving tissue to avoid ghosting artifacts. Lymphatic or venous channels (slow flow) are very bright with presaturation of the spin-echo sequences and less bright with gradient-moment nulling. Areas of slow flow may also be associated with signal loss because of saturation.(40) TOF methods primarily provide morphologic information and only indirectly indicate flow rates.(41) However, TOF has a restricted range, and therefore long vessels are more difficult to visualize.(19) In PC methods, at least two scans in different directions are necessary so that subtraction techniques can be used to suppress background tissue and extract vessel structures. However, in vessels with minimal or absent flow, e.g., thrombus, MRA alone may fail to detect the lesion and a standard MRI may be used as a correlative study.(40)
Although MRA generally provides poorer resolution than does CA, it has greater flexibility in terms of permitting multiplanar projection images of blood vessels containing anatomic detail and quantitative information about flow.(42) A computer algorithm that automatically detects flowing blood provides reconstructed views of vessels while analyzing and displaying flow characteristics.(15)
All TOF techniques involve spin "tagging" in which spins outside an imaged region are labeled ("tagged") and subsequently detected as they enter the imaged region.(22, 26) TOF methods are commonly classified by acquisition mode; e.g., volumetric acquisition, 3D acquisition, or sequential-slice 2D acquisition. (31) Multiple thin-slab acquisition combines both 2D and 3D techniques in which many individual images can be compressed into a single image using a maximum-intensity projection algorithm that selects the maximum signal intensity along a line that cuts through the image space.(39)
In the 2D or 3D methods, the volume of tissue being imaged is subjected to repeated RF pulses, resulting in magnetic saturation and decreased signal intensity. Flowing blood entering this volume, not subjected to magnetic saturation, maintains its signal intensity. This effect, termed flow-related enhancement, is then used to generate images in which flowing blood exhibits high-signal intensity relative to the surrounding stationary tissue.(21)
2D MRA involves an imaging volume limited to very thin slices (approximately 1.5 mm) in which the flowing blood does not become saturated. Sequential slices are then imaged for the area of interest. (43) The major advantages of 2D techniques are the ability to image a large volume in a relatively short time, sensitivity to a wide range of velocities, and ability to obtain significant data even with significant patient motion.(42) Disadvantages include poor spatial resolution and significant signal loss and artifact production from vessels with complex flow.(44) 2D techniques are frequently applied for imaging venous flow in the head, chest, abdomen, and pelvis.(40)
3D MRA uses a relatively thick slice (approximately 4 cm) as the imaging volume, in which the inflowing blood remains for a relatively longer period of time and may become saturated, making 3D MRA relatively insensitive to slow-flowing blood. However, 3D imaging permits sections of less than 1 mm to be obtained in very brief scanning times. 3D MRA offers high spatial resolution with reduced signal loss, making it an excellent technique for imaging fast flow such as seen in arteriovenous malformations.(45) Another advantage of 3D techniques is minimal signal loss from vessels with complex flow. Disadvantages include the fact that saturation effects limit its usefulness in examination of small anatomic regions.(42) 3D techniques are commonly applied for rapid arterial flow generally seen in the head, neck, aorta, and renal arteries.(40)
MRA images may be computer postprocessed, by a variety of techniques, to create multiple-projection images (without additional imaging) in which the vasculature is displayed in a manner similar to that of CA, rather than transverse sections as displayed in MRI.(21, 44, 46)
Although no single 2D image can portray all the information acquired by 3D techniques, postprocessing can yield images of most of the information in the volume image by "compressing" the 3D data onto a 2D image similar to that obtained by CA. However, this postprocessing does not always eliminate the need to inspect individual-slice data from suspicious regions of interest.(47)
Spatial resolution depends on the volume element for a given slice thickness (voxel), and this minimum usable size is largest for 2D TOF and smallest for 3D techniques. Therefore, achievable spatial resolution is best for 3D and limited by slice thinkness for 2D techniques.(42)
PC angiography represents a groups of MRA techniques in which velocity and hemodynamic effects (physiologic information) are incorporated into the vascular image.(18) PC is only sensitive to motion, thereby having better suppression of stationary tissue than is possible with TOF imaging.
PC imaging relies on the concept of predictable phase shifts generated by flowing blood in a magnetic-field gradient.(40, 48) Using this method, the proportional relationship between signal intensity and velocity permits direct velocity measurement associated with marked suppression of background tissue. However, this requires multiple acquisitions and therefore much longer imaging times than those required for TOF methods, especially if only morphologic information is sought.(21, 26) PC methods are associated with more flow-artifact problems and signal loss than TOF methods, but offer unique opportunities for flow imaging.(26, 49) PC techniques have been further refined to produce quantitative measurements in a projected angiographic format to measure both venous and arterial flow at any point in the cardiac cycle without using contrast media.(13) As with TOF methods, computer postprocessing of MRA data is generally used to create projection angiograms in a format similar to that of CA, but which can be viewed in multiple directions.(25, 43, 50)
Essentially, all MRA techniques use variations of 2D or 3D TOF or PC techniques to produce a flow-sensitive image by gating of pulsatile flow and obtain background suppression by postprocessing or subtraction techniques. (51, 52) The original MRI technology, which relied on conventional spin-echo pulse sequencing, had significant limitations for assessing blood vessels relating to the production of complex signal patterns.(53) Newer techniques, which may be applied singly or combined, e.g., RF presaturation flow compensation, gradient-echo pulse sequence, and rapid-line scan angiography, permit the demonstration of vessel patency, flow velocity, and volume and direction of flow.(54)
The hazards and limitations associated with CA have led to the development of competing technologies applied to the evaluation of the vascular system. Ideally, the detection of either an obvious lesion or a normal vascular segment by noninvasive means would be the examination of choice.(55)
The rationale for the use of MRA is predicated on the fact that it involves no exposure to ionizing radiation, is repeatable and relatively rapid, avoids the risks associated with CA (i.e., vascular damage, ionizing radiation, neurologic complications, and systemic reactions to contrast agents), and may be used in an outpatient setting.(41,48,56-62) In addition, MRA can be reviewed retrospectively with little operator dependence, as is seen with ultrasonogaphy, and is not affected by factors that may compromise the use of CA, e.g., poor renal function, coagulopathies, lack of femoral pulses, or the inability to properly advance the catheter because of severe generalized atherosclerotic disease.(63-65)
Proponents claim that MRA can yield rapid and reproducible vascular images with excellent contrast in a manner that correlates well with morphologic results obtained using CA and can readily provide hemodynamic information not available by CA. MRA is easily combined with MRI to yield information on parenchymal morphology as well as vascular dynamics.(25,66)
The literature search encompassed all English-language journal articles and textbooks published between January 1985 and December 1993 available through the search capabilities of the National Library of Medicine. The key words used in the search were "MRI angiography" and "MRA." The search was expanded to include additional references cited in the reviewed articles and textbooks. Additional information was obtained via a Federal Register announcement of this assessment, (66). solicitation of information from professional societies and organizations having interest and/or experience with this technology, preprints of papers submitted for publication, and published abstracts. This review will emphasize those published studies comparing MRA with CA applied to various anatomic sites.
The safety of MRI was addressed in a 1985 Public Health Service Assessment, which concluded that MRI is without demonstrable adverse biologic effects when commercially marketed systems are operated within their specified parameters.(67) However, caution must be exercised with patients having metallic surgical clips or prostheses because of possible adverse effects of MRI-generated magnetic and electromagnetic fields on such items. MRA is contraindicated in patients with pacemakers. In the current "guidance" from the Food and Drug Administration's Center for Devices and Radiological Health (personal communication), the scanning of the fetus or infant is treated as a warning. Claustrophobic anxiety related to long imaging times for patients enclosed within the magnetic bore has been a problem in a small percentage of patients. These cited issues apply equally to the application of MRA. The preliminary and occasional intravenous use of the paramagnetic contrast agent, gadolinium, in MRA has been associated with approximately 1 percent of minor general and renal adverse effects in Phases III-IV studies involving more than 13,000 patients.(68) The observed adverse events were comparable with those seen after intravenous administration of iodinated nonionic x-ray contrast media.
| Reference | Year | MRA method | No. patients/vessels | Agreement [a] (percent) | Accuracy [b] (percent) |
| Masaryk(69). | 1988 | 2D TOF (1.0-or 1.5-T Siemens Magnetom) | 14/14 | 15 | 15 |
| Wagle(48). | 1989 | 3D TOF and 2D PC (1.5-T GE Signa) | 11/11 | NR | 100 |
| Edeman(70). | 1990 | 2D or 3D TOF (1.5-T Siemens Magnetom) | 17/33 | 39 | 100 |
| Kido(71). | 1991 | 2D PC (1.5-T GE system | 31/60 | 53 | 100 |
| Laster(72). | 1991 | 2D TOF | NR/200+ | 90-100 (stenosis-occlusion) | NR |
| Litt(73). | 1991 | 2D TOF (1.5-T Philips Gyroscan S15) | 50/94 | 63 | 100 |
| Masaryk(74). | 1991 | 3D TOF (1.5-T Siemens Magnetom) | 38/75 | 98 | 87 |
| Mattle(75). | 1991 | 2D TOF (1.5-T Siemens) | 20/39 | 92 | 100 |
| Wilkerson(76). | 1991 | 3D TOF (1.5-T GE Signa) | 13/26 | 92 | 100 |
| Furuya(77). | 1992 | 3D TOF (1.5-T system) | 22/26 | 96 | 100 |
| Pavone(78). | 1992 | 2D TOF (0.2-T Hitachi) | 28/54 | 72 | 90 |
| Riles(79). | 1992 | 2D TOF (1.5-T Philips Gyroscan S15) | 41/75 | 52 | 100 |
| Wesbey(80). | 1992 | 2D TOF (1.5-T GE system) | 25/37 | 86 | 97 |
| Anson(81). | 1993 | 2D and 3D TOF (1.5-T Siemens Magnetom) | 20/40 | 95 | 100 |
| Blatter(82). | 1993 | 3D TOF (1.5-T GE Signal) | 51/102 | 89 | 99 |
| Freeman(65). | 1993 | Unspecified | 17/36 | 79 | 92 |
[a] Concurrence in depicting lesion size vs. overestimation or underestimation.
[b] Identification of normal or abnormal vessel.
Abbreviations: NR = not reported; T = tesla.
| Reference | Year | MRA method | No. patients | Study vessel(s) | Agreement [a] (percent) | Accuracy [b] (percent) |
| Edelman(83). | 1989 | 3D TOF (1.5-T Siemens) | 10 | AVMs | 100 | 100 |
| DeMarco(84). | 1990 | 2D TOF (1.5-T GE Signa, 11 patients; 0.5-T Diasonics, 1 patient) | 12 | Dural arteriovenous fistulae | 63 | 100 |
| Marchal(85). | 1990 | 3D TOF (1.5-T Siemens Magnetom) | 17 | AVMs, venous angiomas | 18 | 94 |
| Ross(86). | 1990 | 3D TOF cine and cine plus spin-echo studies (1.5-T Siemens Magnetom) | 19 | Aneurysms | 84 | 89 |
| Blatter(87). | 1991 | Multiple overlapping thin-slab acquisition TOF (1.5-T GE Signa) | 11 | Intracerebral arteries | NR | 89 |
| Nadel(88). | 1991 | 2D PC (1.O-T Siemens Magnetom) | 14 | Varied abnormalities | NR | 100 |
| Nussel(89). | 1991 | 3D PC (1.5-T GE system) | 10 | AVMs | 70 | 100 |
| Applegate(90). | 1992 | 2D PC (1.5-T system) | 17 | Varied abnormalities | NR | 94 |
| Schuierer(91). | 1992 | 3D TOF (1.0 or 1.5-T system) | 18 | Aneurysms | 59 | 83 |
| Awad(92). | 1993 | 2D and 3D TOF (1.5-T Siemens Magnetom) | 31 | Aneurysms, AVMs | 45 | 94 |
| Ostertun(93). | 1993 | 2D TOF (1.5-T and 0.5-T units) | 12 | Venous anomalies | 25 | 81 |
| Petereit(94). | 1993 | 2D PC and 3D TOF (unspecified system) | 20 | AVMs | 91 | 100 |
[a] Concurrence in depicting lesion size vs. overestimation or underestimation.
[b] Identification of normal or abnormal vessel.
Abbreviations: AVMs = arteriovenous malformations; NR = not reported; T = tesla.
| Reference | Year | MRA method | No. patients | Study vessel(s) | Agreement [a] (percent) | Accuracy [b] (percent) |
| Amparo(58). | 1985 | 2D TOF (0.35-T Diasonics MT/S) | 10 | Aortic dissections | 100 | 100 |
| Hricak(59). | 1985 | 2D Fourier transform (0.35-T Diasonics MT/S) | 11 | Abdominal venous thromboses | NR | 91 |
| Valk(37). | 1985 | 2D TOF (0.35-T Diasonics MT/S) | 12 | Abdominal aorta | 93 | 100 |
| Kim(97). | 1990 | 2D TOF (1.5-T Siemens Magnetom) | 25 |
|
|
|
| Debatin(98). | 1991 | 2D TOF and PC (1.5-T GE Signa) | 24 | Proximal renal artery | 91 | 100 |
| Kent(99). | 1991 | 2D TOF (1.5-T Siemens) | 37 | Renal artery | 91 | 95 |
| Linden(100). | 1991 | 2D TOF (1.5-T system) | 12 | Renal artery | 50 | 100 |
| Vock(96). | 1991 | 3D PC (1.5-T GE Signa) | 17 | Renal arteries and portohepatic collaterals | NR | 94 |
| Gedroyc(101). | 1992 | 3D PC (1.5-T GE Signa) | 48 | Renal artery (posttransplant) | 92 | 100 |
| Durham(102). | 1993 | 2D TOF (1.5-T GE system) | 28 |
|
|
|
| Smith(103). | 1993 | 3D TOF (1.5-T Siemens Magnetom 63) | 20 |
|
|
|
[a] Concurrence in depicting lesion size vs. overestimation or underestimation.
[b] Identification of normal or abnormal vessel. Abbreviations: NR = not reported; T = tesla.
The effects of both cardiac and respiratory motion have been regarded as serious handicaps to the application of MRA in evaluating thoracic vasculature. However, recent technical advances, including echoplanar imaging (a method that collects 2D information from a single RF excitation) and steady-state methods (which use faster imaging to overcome the effects of motion), have served to overcome these problems.(104,105)
| Reference | Year | MRA method | No. patients | Study vessel(s) | Agreement [a] (percent) | Accuracy [b] (percent) |
| Dinsmore(106). | 1986 | 2D Fourier technique (0.6-T Technicare) | 6 | Thoracic aorta | 97 | 100 |
| Kauczor(107). | 1991 | FLASH 2D TOF (1.5-T Siemens Magnetom SP) | 6 | Intrathoracic vasculature | NR | 100 |
| Manning(108). | 1993 | 2D TOF (gradient echo) (1.5-T Siemens Magnetom SP) | 39 | Coronary arteries | 89 | 91 |
[a] Concurrence in depicting lesion size vs.overestimation or underestimation.
[b] Identification of normal or abnormal vessel. Abbreviation: NR = not reported.
| Reference | Year | MRA method | No. patients | Study vessel(s) | Agreement [a] (percent) | Accuracy [b] (percent) |
| Mulligan(110). | 1991 | 2D TOF (1.5-T Philips Gyroscan) | 12 | Iliac, femoral, and poplitel arteries | 71 | 67 |
| Gehl(111). | 1991 | 3D TOF (1.5-T Siemens Magnetom) | 16 | Hemodialysis fistulae | 13 | 100 |
| Owen(112). | 1992 | 2D TOF (1.5-T GE Signa) | 73 | Infrarenal aorta to feet | 82 [c] | 100 |
| Hertz(113). | 1993 | 2D TOF (1.5-T GE Signa) | 19 | Distal aorta, iliac, femoral, popliteal, and crural arteries | 83 | 100 |
| Spritzer(114). | 1993 | 2D gradient-recalled echo (1.5-T GE Signa) | 54 | Inferior vena cava distal to popliteal veins | NR | 96 |
| Yucel(115). | 1993 | 2D-TOF (1.5-T GE Signa) | 25 | Distal aorta through popliteal trifurcation | 84 | 100 |
[a] Concurrence in depicting lesion size vs. overestimation or underestimation.
[b] Identification of normal or abnormal vessel.
[c] MRA demonstrated additional patent vessel segments not seen by CA in 9 patients.
Abbreviation: NR = not reported; T = tesla.
Although CA remains the diagnostic standard for most clinical applications, it is not unfailingly accurate, not always practical or feasible, and competing noninvasive technologies (e.g., MRA, ultrasound) are rapidly making inroads into its domain.(8,96,116-119) In some clinical situations, the application of MRA for the identification of vascular pathology is purported by some investigators to hold the potential for a complete noninvasive evaluation.(40,120-122). This includes the ability to distinguish dissections; slow flow from thrombus; occlusions from patent vessels; significant from insignificant degrees of stenosis; and the general assessment of direction of flow and velocity, including inflow and outflow vessels as roadmaps preceding surgery. In addition, it is claimed that MRA has the potential of being used as a followup for most, if not all, angiographic studies.(16) Clinically, the value of flow quantification is largely unproven, and MRA techniques are used to indicate direction of flow or assess collateral flow patterns.(40) However, none of the available MRA methods can accurately measure the full range of flow velocities, and this difficulty has contributed to the lack of widespread acceptance of MRA for routine clinical applications.(29) Because CA is the reference standard for evaluating the vascular system in most clinical situations, MRA continues to be clinically evaluated against angiography, and this research is mostly limited to larger centers using sophisticated MR equipment capable of providing high field strengths.(122,123)
Although it is expected that further refinements in techniques will result in MRA supplanting many current CA procedures, at present, controversy continues and opinions abound concerning both the current use and future applications of MRA; and, as yet, there are no prospective randomized trials comparing MRA with CA in evaluating any vascular region.(1,46,54,79,124,125) A major limitation of MRA compared with CA is its significantly poorer spatial resolution (2D image sharpness) (16,21,107,126-128) Improvements in spatial resolution and decreasing image acquisition time are primary goals of current MRA research.(128,129) Other limitations and drawbacks of MRA include prohibitions concerning patients with pacemakers or other metallic implants, motion artifacts, low cardiac ejection fractions, and distorted images related to high-velocity flow and its turbulence in areas of severe stenosis.(78,88,111,130).
The examination time and cost of MRA is often less than that of CA; however, its use as an adjunct to CA would obviously increase the total cost.(76,112,130-132)
The largest experience of MRA has been in evaluating the carotid and intracerebral arterial circulation, areas relatively free from cardiac and respiratory (38,64,77). These indications have been gaining wider acceptance and include MRA use in clarifying equivocal findings of other imaging techniques and willingness of a few surgeons to perform surgery in selected cases on the basis of MRA alone or in addition to Doppler scanning.(21,63,80,81,121,133,134).
A review of the reports listed in the tables indicated that the studies were not comparable and in general had small patient samples (nonuniformly selected), with the images evaluated in some cases with or without prior knowledge of the CA results. In almost all cases, imaging was performed using state-of-the-art technology; however, imaging techniques were not uniform. In some cases, MR angiograms and those of the reference test (CA) were excluded from analysis because of nonvisualization or poor quality of the images.(71,79,84,135) These factors engendered difficulties for the interpretation of these noncomparable studies.
In attempting to evaluate the quality of evidence of MRA applied to the extracranial carotid arteries in the 16 reports listed in Table 1, it was noted that in the two earliest studies (1988, 1989) patients selected for MRA were those with angiographically documented disease, and the MRA-CA comparisons were unblinded.
The report of the largest experience with more than 100 patients(72). appeared in abstract form; neither indicated how patients were selected for examination nor whether the MRA-CA comparisons were blinded.
Of the 20 patients reported by Mattle, (75). seven were classified as asymptomatic, although one might presume they were selected because of an abnormal duplex scan; the reason for their selection is unstated; and the method of comparison is unspecified.
Twelve of the 16 reports did specify blinded comparisons of MRA-CA. Except for the reports by Wesbey, (80). in which the selection of patients is unspecified, and in the reports by Riles(79). and Furuya, (77). which included asymptomatic patients with disease documented by duplex scanning, all patients selected for evaluation were symptomatic.
In the studies of head and neck vasculature, excepting an earlier (1988) study, MRA clearly and correctly identified the arteries as patent, stenotic, or occluded in the majority of cases, although overestimation of the degree of stenosis was frequently noted and accounted for most of the lower percentages of agreement with CA. The presence of artifacts was the principal cause for the failure of MRA to detect vascular disease.
In evaluating the intracerebral vasculature, the demonstrated ability of MRA to detect pathology ranged from 81-100 percent. However, its agreement with CA in delineating the degree of pathology was less than that in evaluating the carotid arteries, with a mean of only 62 percent in the nine of 12 studies in which agreement was reported. The total number of patients in these 12 reported studies was 191, with the largest involving 30 patients.(92) All the comparisons of MRA and angiography were made on patients with angiographically documented disease; however, in only two of the 12 reported studies.(83,86) were the comparisons made by readers unblinded to the angiographic results.
It has been stated that presurgical and postsurgical measurements of carotid blood-flow volume are useful in patients with such conditions as occlusions and arteriovenous malformations (AVMs) (93) However, whether MRA has an advantage over ultrasound techniques is questionable. Carotid ulcerations may not be detected because of the relative stasis of blood failing to produce adequate blood-flow contrast.(6,87,136) For most clinicians, duplex ultrasound scanning, because of its relative ease, lower cost, and safety, remains the preferred technique for differentiating normal and abnormal carotid vasculature, and CA continues to be the definitive diagnostic and preoperative study in carotid disease.(41,57,58,63,89) Sonography is limited in the presence of large calcified plaques by its inability to demonstrate the intracranial portion of the carotid artery and its dependence on the skill of the operator.(137). MRA can selectively image most of the proximal intracranial and extracranial arteries and veins and provide morphologic information closely related to their hemodynamics.(38,138) Combined with MRI, it can survey the brain parenchyma and the entire carotid and vertebral territory and provide a noninvasive examination in situations such as sickle cell patients with focal neurologic dysfunction who are at high risk for stroke.(21,139-141) MRA is rapidly acceptance in the radiology community as a noninvasive diagnostic technique for patients with suspected occlusions or vascular malformations, especially aneurysms, and the combined use of MRI with MRA has been used to evaluate patients with suspected intracerebral arteriovenous malformations and cerebral venous occlusions (where MRA can provide a 3D anatomic display not available from MRI alone).(64,129,138,142). However, CA continues to be the definitive preoperative study. MRA has been used as the primary or sole morphologic study in patients with cerebrovascular symptoms. MRA has also been used to guide stereotactic radiation therapy for cerebral angiomas.(143) A significant limitation of intracerebral MRA continues to be in areas of slow flow, such as giant aneurysm and venous angiomas that are poorly visualized due to saturation effects of blood residing too long in the volume of excitation.(10) In general, CA is required in the evaluation of patients with suspected aneurysms, vascular malformations, and arteritis and remains the standard technique to detect, localize, and quantify occlusive cerebrovascular lesions.(138) On the basis of the results seen in Table 1, one might conclude that the ability of MRA to detect the presence or absence of vascular disease in the region of the head and neck is quite acceptable. A recent review of medical progress in neurology stated that MRA applied to cerebral arteries is most useful in screening patients for CA and serial monitoring of vascular abnormalities.(144) If these prove to be useful applications, the costs involved in managing these patients would be reduced.
The application of MRA in the peripheral, abdominal, and thoracic vasculature has resulted in variable success. This has been attributed to different imaging requirements (e.g., large field of view), different patterns of blood flow, the suppression of overlapping vasculature and stationary tissues, and technical difficulties related to motion artifacts and to the use of body rather than surface coils.(120,121)
A recent review of MRA in assessing diseases of the thoracic aorta concluded that it is the technique of choice for providing quantitative information in the evaluation of aneurysms, dissections, aortitis, trauma, and valvular disease. It is particularly useful in evaluating pseudoaneurysms involving aortocoronary bypass grafts and in aortic annular pseudoaneurysms associated with bacterial endocarditis.(145) However, this opinion is primarily based on preliminary reports of the application of MRA to acquired disease of the thoracic aorta.
MRA for the imaging of coronary arteries has been relatively unsuccessful primarily due to such problems as respiratory and cardiac motion artifacts, the small and pulsatile nature of the vessels, irregular flow patterns, and interference from blood in other cardiac chambers and vessels.(46,96,146-149) Coronary MRA is also relatively dependent on the ability to hold one's breath for 12-18 seconds and the presence of a regular heart rate and rhythm.(108) In addition, the spatial resolution of MRA in this area does not currently match that of CA (the current standard for evaluating coronary arteries), and partial-volume effects and signal dropouts caused by turbulence preclude its use for the quantitative assessment of stenoses.(5,108) The relatively larger size of coronary bypass grafts has permitted MRA to be used in assessing their patency.(148) More recent techniques involving cardiac gating and ultrafast echo-planar MRA appear to offer potential for the clinical utility of depicting the coronary vasculature and offering a noninvasive identification of the degree of coronary artery disease.(108,131,150-152) Preliminary data have shown that MRA, when compared with CA, can achieve sensitivity and specificity on the order of 90 percent.(108) If verified by others, this level of accuracy would make MRA a competitive procedure for coronary artery imaging. A recent editorial suggested that MRA applied to coronary artery disease is best reserved for screening rather than treatment planning or monitoring patients with known coronary artery disease, until such time as the resolution of MRA is improved and studies are performed comparing its clinical utility and cost effectiveness with other imaging techniques.(152)
The established MRA techniques currently used for the cerebral, neck, abdominal, and peripheral vascular studies have not been successful in generating adequate angiograms in the pulmonary system, primarily because of motion artifacts, magnetic artifacts related to air-tissue interfaces, and the complex overlapping of pulmonary arteries and veins.(153)
A relatively new application of MRA is MR venography, and experience with this technology in the thorax and abdomen is growing, with the bulk of the examinations being obtained with TOF techniques. The complex nature of the thoracic veins and the limited acoustic access of the region make it difficult for ultrasound or computed tomography to visualize these veins.(154) Contrast venography, as the reference standard, is also associated with limitations relating to venous access, flow effects, and the use of contrast media. In a study of 55 patients comparing MRA with contrast venography applied to the thorax, access was not available in five of the patients, so that conventional venography or computed tomography was impossible.(154)
The clinical application of MRA to the study of the abdominal veins is also seen as a reasonable alternative to the widely used ultrasound, computed tomography, and contrast angiography examinations of the abdomen.(154,155) The presence of bowel gas and unfavorable acoustic angles serve to undermine ultrasound examinations in this area. Computed tomography requires large amounts of contrast media with its associated risks, and contrast angiography may not be effective in patients with abnormal venous anatomy.(155) MR venography may be routinely applied to patients with portal disease or suspected occlusions of the interior vena cava, iliac, or renal veins.(155)
For imaging the larger abdominal vessels, MRA is diagnostically reliable in detecting aortic occlusions and aneurysms and stenoses of the proximal renal artery.(156-158). However, its spatial resolution is insufficient in areas of slow flow and the smaller abdominal vasculature. Renal arteries are often difficult to image using MRA because of motion artifacts (cardiac, respiratory, and bowel), complex flow patterns, and complex flow directions. The conventional identification of renovascular disease in patients with hypertension has been invasive and has required the use of contrast agents, thus precluding its suitability for routine diagnostic examinations.(159) A recent review of newer tests for the diagnosis of renovascular disease confirmed the theoretical advantage of MRA in the diagnostic of RAS, but also noted the lack of prospective studies investigating its utility for this purpose and concluded that additional studies are necessary to establish diagnostic criteria.(160) A report from the Mayo Clinic has also noted the potential of MRA as the most effective noninvasive test for RAS.(161) In some studies, the reported visualization of renal artery origins ranged from approximately 30-50 percent of cases.(162,163) Other reports indicate specificity and sensitivity in detecting significant RAS in the range of 90-100 percent.(3,96,97,99) Although not directly compared with other techniques, MRA renal blood-flow measurements in normal volunteers compared well with published measurements in other normal subjects.(164,165) MRA is being increasingly applied to the study of the portal venous system, especially in patients with portal hypertension, and in the preoperative assessment for liver transplantation.(55,96,166)
In reviewing the reports in Table 5, the results in the Owen study(112). involved 73 patients with severe symptomatic peripheral vascular disease, and the MRA-CA comparisons were blinded. The four other studies of peripheral arterial vascular disease involved small numbers of patients (total: 66). The very poor agreement (13 percent) between MRA and CA in the unblinded comparisons reported by Gehl(111). was related to the motion artifacts associated with the wide range of velocities of both the arterial and venous blood flow, leading to signal voids. The Spritzer study(114). involved MRA and contrast venography in 54 patients with suspected deep venous thrombosis. Venography results were known at the time of MRA interpretation in five of the 54 cases.
A report of the Ad Hoc Committee of Western Vascular Society has concluded that there is insufficient evidence that MRA is or will be superior to either CA or well-done ultrasound duplex studies.(167)
Although CA is regarded as the standard technique for evaluating peripheral artery disease, it fails to identify distal vessels suitable for reconstructive surgery in up to 70 percent of cases with severe stenoses.(168) The superior ability of MRA to detect patent distal vessels has led to its use, by some investigators, as the sole preoperative study.(113)
In response to the Federal Register notice of this assessment(66). and the solicitation of information and opinions from organizations and institutions involved with MRA, the Office of Health Technology Assessment (OHTA) has received the following input:
Duke University Medical Center (November 1992) believes that MRA has demonstrated its potential in delineating the anatomy and pathology of the pulmonary vasculature and in accurately diagnosing deep venous thrombosis in both the pelvis and calf.
University of Wisconsin-Madison Medical School (December 1992) has had experience with more than 2,000 MRAs and found it to be a valuable clinical tool. They have begun performing carotid endarterectomy in symptomatic patients without preoperative CA when there is an exact correlation between MRA and duplex imaging. The combined costs for this preoperative workup is approximately $2,100 vs. approximately $5,500 when the workup includes CA.
The Cleveland Clinic Foundation (January 1993) believes that MRA is probably at a stage whereby it can effectively act as a screening modality for carotid stenosis in lieu of Doppler ultrasound and has the potential for challenging CA as the presurgical examination of choice. Preliminary studies suggest MRA can act as an effective screen for intracranial atherosclerotic disease and berry aneurysms. It is inappropriate to suggest the use of MRA for the detection of aneurysms in patients with documented subarachnoid hemorrhage.
The American College of Radiology (January 1993) has listed the following indications for MRA: evaluation of suspected disorders of the intracranial and extracranial vasculature, including cervicocranial arterial dissection, dural sinus thrombosis, cerebral arteriovenous malformations, carotid bifurcation atherosclerosis, vascular status following extracorporeal membrane oxygenation, intracranial aneurysm, sickle cell anemia with neurologic complications, and carotid body tumors.
Specific indications for thoracic application include dissection of the thoracic aorta; aneurysms of the aorta, its branches, and the pulmonary arteries; differentiation of aneurysms and masses and definition of the relationship of masses to nearby vascular structures; identification of congenital abnormalities of the aorta; and evaluation of the superior vena cava syndrome or unilateral upper-extremity edema.
MRA routinely demonstrates the major abdominal vessels and can detect stenosis of the main renal arteries and has been successful in the detection of renal vein thrombosis and evaluating patients with portal hypertension. Other indications include detection of splenic vein thrombosis, portal vein invasion by hepatic neoplasms, dissection or aneurysms of the abdominal aorta, aorta-occlusive disease, evaluation of vascular tumors, patients in whom iodinated contrast is contraindicated, evaluation of patency of portocaval shunts, and evaluation of abdominal aortic grafts.
Indications for MRA in the extremities include detection of hemodynamically significant stenoses in patients suspected of ischemia, detection of pseudoaneurysms in patients' postbypass grafting or penetrating trauma, determination of the relationship of masses to nearby vasculature, and the evaluation of the patency of vascular grafts.
The Society for Vascular Surgery (January 1993) has stated that no randomized trials have been performed to evaluate the effectiveness of MRA. MRA devices have been primarily designed to compete with CA. Measurement of blood flow and study of vessel wall morphology should be regarded as highly experimental. The concomitant use of MRA with MRI for imaging the brain is at least as good as computed tomography of the head (and perhaps better). However, if one is looking at MRA alone for imaging the head and neck, there is no evidence that it is comparable to or better than duplex scanning or contrast angiography. The results of MRA used to visualize distal arteries of the leg and foot have been promising, but no comparative studies have confirmed its value.
The Pittsburgh NMR Institute (January 1993) has found MRA to be reproducible and reliable in evaluations of intracranial arterial malformations (pre-and postinterventions), intracranial aneurysms detection and followup, common carotid bifurcation, and questions regarding arterial dissection. MRA is routinely used for surgical planning for arterial feeder supply and mapping of venous drainage for various intracranial and spinal neoplastic processes. MRA is used to assist in the care of patients undergoing gamma knife therapy.
The American College of Surgeons (January 1993) offers the following opinions concerning MRA:
The large MR equipment does not measure blood-flow rate (mL/min), although it does show flow and can be used to estimate velocity. Smaller permanent MR equipment developed specifically for blood-flow scanning does have the capability to measure blood-flow in mL/min accurately.
High-quality MRA, particularly with 3D capabilities, is probably the method of choice to study intracranial vessels.
MRA of the extracranial carotid arteries is still significantly less useful than either conventional contrast arteriography or duplex ultrasound scanning.
In examination of the venous circulation, MRA may be useful in the study of abdominal veins but is not competitive with duplex ultrasound for lower-extremity venous evaluation.
MRA is a more accurate evaluation of the hemo-dynamic effects of percutaneous transluminal angioplasty than are standard pressure, Doppler velocity, or plethysmographic techniques, which have been shown to be frequently inadequate.
The research application of MRA to measure lower-extremity hemodynamic response to medications or exercise programs offers a significant advantage over traditional physiologic testing.
Additional uses of MRA, particularly the application of small, permanent magnet units which can measure lower-extremity blood flow, include: assessment of whether ulcers will heal without revascularization, evaluation of the degree of ischemia in diabetics with vessel calcification, determining the level of amputation, assessing arterial insufficiency in Raynaud's phenomenon, and research evaluation of agents affecting vasoconstriction or blood viscosity.
Intracerebral vessels: MRA is currently useful for screening patients for intracranial aneurysms, particularly of the Circle of Willis. Conventional angiography is still needed for patients for whom surgery is indicated. MRA is also useful for estimating intracerebral blood flow and in planning for treatment of some intracranial tumors and vascular malformations.
Cervical carotid arteries: Although clinical studies are ongoing, MRA is currently no more sensitive or specific than duplex Doppler ultrasound with or without color-flow analysis. MRA is more expensive than duplex technology, and screening with duplex scanning and conventional angiography for patients who are candidates for carotid surgery remain the diagnostic standards.
Peripheral arteries: MRA may provide useful imaging but is probably no more sensitive or specific than duplex ultrasonic scanning with color-flow assessment and CA or digital arterial angiography for patients who are candidates for operations. At present, MRA will not likely supplant contrast angiography in terms of lower-extremity occlusive disease evaluation and management, although it is certainly possible to carry out lower-extremity vascular reconstruction without angiograms in selected circumstances.
Aorta and visceral arteries: Adequate visualization of renal or mesenteric vessels is difficult, and the current state of MRA is not particularly helpful in these areas. Gated MRI imaging appears to be more useful than MRA. However, MRA may eventually be developed to the point whereby it is useful for screening for renal or visceral artery disease, and the technique may also eventually be useful in lieu of CA for patients with abdominal aortic aneurysms.
Peripheral venous disease: MRA does not offer any value over duplex scanning or conventional venography for deep vein thrombosis.
Coronary arteries: Currently, MRA is not clinically useful for evaluation of coronary artery disease. Further development of the technology may eventually prove useful, however.
The Mayo Clinic (March 1993) states that MRA has totally replaced intravenous digital subtraction angiography in their neuroangiographic practice. Questions of dural sinus thrombosis and patency of carotid arteries following endarterectomy are answered by MRA. MRA equals Doppler color-flow ultrasound in the evaluation of carotid stenosis.
Cine PC MRA will reliably demonstrate that patency of third ventriculostomies. MRA is used as followup to AVMs after radiosurgery. MRA is not a replacement for all CA. A patient presenting with an acute subarachnoid hemorrhage must have a selective catheterization angiogram. Although MRA equaled ultrasound for evaluating carotid stenosis, it is more expensive and less readily available. Ultrasound is used as a screen for carotid disease, and patients are then subjected to CA if surgery is contemplated. MRA is reserved only for those patients with difficult diagnostic situations or equivocal ultrasound results. MRA is superior to computed tomography for evaluating patients with possible vertebrobasilar ischemia or infarction.
MRA has consistently been able to detect intracranial aneurysms >0.3 mm in size, and offers a noninvasive method to follow aneurysms in patients with autosomal-dominant polycystic kidney disease. In patients with arterial venous malformations, MRA offers unique information that can be complementary to that provided by CA. Cine MRA techniques have been applied to candidate cerebrospinal fluid flow in patients suspected of normal-pressure hydrocephalus.
The American College of Cardiology (April 1993) has indicated the following clinically accepted uses of MRA:
Extracranial carotid and vertebral arteries for detection and assessment of severity of atherosclerosis, dissection, and other lumenal abnormalities.
Intracranial arteries and veins for identification of aneurysms, arteriovenous malfunctions, and relationships of arteries and veins to masses.
Thoracic aorta and major branches (excluding coronary arteries) --evaluation of intrinsic and extrinsic pathology, e.g., atherosclerotic and inflammatory diseases, dissections, aneurysms, coarctation, and trauma. Identification of arteriovenous fistula and malformations of thoracic aortic branches, i.e., spinal arteriovenous malformations.
Abdominal arteries and veins, including the aorta and visceral branches; the inferior vena cava and primary branches, as well as the portal vein and splenic and superior mesenteric veins.
Renal arteries --detection of stenosis of proximal renal arteries.
Peripheral arterial obstructive disease and evaluation of peripheral runoff arteries. Evaluation of peripheral arterial aneurysms and malformations.
Detection of pelvic and peripheral venous thrombosis.
MRA is indicated in all patients with suspicion of any of the above-listed abnormalities, except those patients in whom MRI is specifically contraindicated. These contraindications are:
Cardiac pacemakers.
Intracranial ferromagnetic aneurysm clip.
Implanted electromagnetic devices such as neurostimulators, bone-growth simulators, or cochlear implants.
Intraocular ferromagnetic foreign bodies.
Automatic cardiac defibrillative devices.
Pregnant women, unless another nonradiation test will not provide urgently needed diagnostic data.
The University of Pennsylvania Medical Center (June 1993) believes that MR venography provides accurate imaging of the central and peripheral veins which is highly sensitive for the detection of deep venous thrombosis and has the potential to replace both CA and venography in evaluating the peripheral vasculature. MRA can also provide postprocedural imaging of angioplasty and arthrectomy sites to reveal vessel wall abnormalities not appreciated by CA.
The Stanford University Department of Radiology (October 1993) on behalf of the MRA Subcommittee of the Society of Magnetic Resonance Imaging provided the following data and opinion. In a study of 139 carotid bifurcations evaluated by MRA and CA, there was nearly equal accuracy at detecting stenoses over the range of 0-100 percent. The sensitivity for detection of>90 percent stenosis was 95.5 percent, and the specificity was 89.4 percent.
MRA has proven useful in intracranial, carotid, pulmonary, liver, renal, and peripheral vessels. The ability to provide both morphologic and functional images impacts clinical diagnosis and treatment.
The Food and Drug Administration (FDA; October 1992) has advised OHTA that they have received and are continuing to receive 510(k) notifications for software additions as normal expansions of various scan sequences used for MRI "flow imaging" as a generic claim. The FDA requires that the labeling contain the warning that MRA is not intended to replace CA. In addition, the labeling for MR diagnostic devices includes the contraindication in the presence of intracranial aneurysmal clips, unless the clip is not magnetically active, and in the presence of active electrical, magnetic, or mechanical implants whose operation may be adversely affected by magnetic and/or electromagnetic fields produced by the MR device. Future claims for MRA will have to be supported by a good clinical study demonstrating the effectiveness claimed.
NIH has advised OHTA that MRA can provide high-quality images that are becoming acceptable for diagnoses or evaluation, in some cases giving information not available by CA methods, including data on flow. MRA has the potential to replace invasive, hospital-based procedures. In several areas, MRA meets or exceeds all the essential requirements of arteriography. Because of discrepancies between MRA and conventional modalities, MRA cannot totally replace CA.
Indications are numerous and continue to expand. Neurovascular studies predominate (carotid bifurcation in atherosclerotic disease, intracranial aneurysms, large-vessel atherosclerotic disease, arteriovenous malformations, and dural sinus occlusions). Portal vein MRA may be used to assess liver transplant candidates. MRA of renal arteries (hypertensive patients or patients with arteriosclerotic disease), deep veins (thrombosis), or pulmonary arteries before lung transplant is used by some teams. Pelvic and lower-extremity MRA studies are under evaluation for lower peripheral arterial-occlusive disease. Preliminary findings on MRA of coronary arteries, regarding sensitivity and specificity in the range of 90 percent as compared with CA, indicate that rapid advances are taking place and that the potential is high for the use of MRA as a clinical and as a screening method for coronary artery imaging within the next several years.
Although MRA appears promising for visualization of the large vessels in particular, the resolution and reliability have not been ascertained such that it is considered a substitute for arteriogram at this time. It does, however, look promising that in several years this may be the case.
Carefully controlled, prospective trials must be conducted to more clearly define the precise roles MRA may play in evaluating patients.
MRA techniques generate contrast between flowing blood and surrounding tissue and provide anatomic images that can be projected in a format similar to that of CA and can also provide physiologic information.
MRA has been extensively applied to the study of head and neck vasculature (especially the carotid bifurcation) with a demonstrated diagnostic accuracy in many patients comparable to that seen with CA. The clinical utility of MRA in evaluating the cerebral vasculature (often combined with MRI) has also been documented for diagnosing patients with suspected intracerebral lesions. However, in the vast majority of cases, CA continues to be the definitive preoperative study.
The MRA evaluation of the thoracic and abdominal aorta and its major branches has also been shown to have a high degree of accuracy. It is being applied in many institutions for detection of aneurysms, occlusions, and stenoses, especially in circumstances in which the use of contrast agents are contraindicated or regarded as high risk.
Reports of MRA of the pulmonary and coronary vasculature represent preliminary experience, with the potential for future clinical utility in the noninvasive evaluation of coronary artery disease.
CA continues to be regarded as the standard technique for evaluating peripheral artery disease; however, the recently demonstrated superior ability of MRA to detect distal patent vessels has led to its use as the sole preoperative study.
The comparisons of conventional and MR venography have been few but suggest that the MR technique has the potential to become the imaging method of choice.
The availability of MRA is often limited to large medical centers, and there is a paucity of information regarding the costs of MRA vs. CA. However, the limited data indicate MRA to be less costly than CA, especially if performed in outpatient facilities. Although MRA may be individually less costly than arteriography, if it were used in "screening," the aggregate costs would likely be much higher due to a lower threshold for the use of MRA related to the relative risks and contraindications of CA.
With regard to the issue of the relative accuracy and agreement between MRA and CA, it should be noted that the clinically acceptable accuracy of MRA in identifying normal or abnormal vasculature may be clinically inadequate when treatment management decisions, such as those applied to carotid endarterectomy, depend on a high degree of precision. To date, MRA results are generally less precise than those obtained with CA methods.
MRA is more clinically useful in large rather than small diameter vessels and continues to be associated with significant technical limitations for some applications, e.g., coronary artery imaging.
On the basis of a growing clinical experience with MRA demonstrating clinical usefulness comparable to that of other imaging techniques (albeit fewer than 1,000 cases directly comparing MRA with CA), MRA appears to be a promising technology for accurately imaging blood flow and blood vessel morphology and potentially replacing competing invasive and noninvasive imaging methods in a variety of clinical situations.
Despite the conspicuous absence of prospective randomized trials demonstrating its clinical effectiveness, MRA is gaining wider acceptance. Although MRA cannot yet be regarded as a standard technique for routine imaging or as a reliable noninvasive alternative to CA in most clinical settings, clinical trials need to be conducted to define more clearly the roles MRA may play for both screening and patient management. MRA can be applied to patients who are poor candidates for standard angiography.
a] MRI can also be based on elements as sodium or phosphorus, but experimental technology is outside the subject area of this assessment.
